RFD: comp.p2p-grid
Rajkumar Buyya
unmoderated group comp.p2p-grid
This is a formal Request For Discussion (RFD) for the creation of
world-wide unmoderated Usenet newsgroup comp.p2p-grid. This is
not a Call for Votes (CFV); you cannot vote at this time. Procedural
details are below.
Newsgroup line:
comp.p2p-grid Peer-to-Peer / Grid Computing.
RATIONALE: comp.p2p-grid
Grid Computing is a new paradigm that enables sharing of geographically
distributed resources such as computers, databases, and scientific
instruments in a secure and consistent manner without any central control
for solving large-scale problems in science, engineering, and commerce.
Grid computing is emerging as the next generation of computing
paradigms. The Grid enables the creation of Virtual Enterprises (VEs)
that can share geographically distributed resources, such as
computers and data sources, to solve, for example, large-scale
problems in science, engineering, and commerce. Already international
forums like:
* Global Grid Forum: <http://www.gridforum.org>
* Peer to Peer Computing WG: <http://www.p2pwg.org>
have evolved to create standards and protocols for inter-operability
between heterogeneous systems providing virtual services. Experts
have predicted that Grid is the next revolution in Internet after the
Web. All over the world, there are numerous projects exploring the
use of these technologies. See the following for some details:
* http://www.gridcomputing.com/
* http://www.computer.org/dsonline/gc/index.htm
Already, numerous companies, including IBM, Sun, DSTC, and Microsoft,
are embracing this technology and investing in it heavily. Therefore,
we believe that it is important to create a newsgroup that will
enable discussion that helps in making technologies surrounding the
Grid a reality and consequently accessible to all.
Although over 20 discussion mailing lists operated by individuals or
institutions exist, they are generally meant for discussion of
specific issues (e.g., standards). Another concern is that mailing
lists are likely to generate large volume of email for members;
therefore many prospective participants feel discouraged from
subscribing, do not become members, and do not join important topical
discussions. Instead, many prefer to participate in discussions as
and when they want without getting swamped with emails. Therefore, we
believe having a newsgroup dedicated to Grid computing helps overcome
all these limitations and will encourage discussion and dissemination
without the need of explicit membership.
Strategy for publicising the comp.p2p-grid newsgroup:
The formation of the comp.p2p-grid newsgroup will be publicised through
the following channels (but not limited to):
* IEEE DS Online,
* Global Grid Forum,
* P2P WG,
* Grid Infoware,
* IEEE/ACM conferences:
* CCGRID'xy: <http://www.ccgrid.org/>,
* GRID'xy: <http://www.gridcomputing.org/>,
* Yahoo Group on gridcomputing as part of GridInfoware.
* IEEE Task Force on Cluster Computing (TFCC)
* Newsgroups such as comp.parallel and comp.distributed.
END RATIONALE.
CHARTER: comp.p2p-grid
Appropriate areas of discussion would include (but not limited to):
* Metacomputing, meta-problems,
* Wide area distributed computing,
* Protocols and interfaces for remote access to resources,
* Local and distributed resource management,
* Security, authentication, and authorization policies,
* Resource brokers,
* Scheduling policies,
* Computational economy,
* Quality of Services,
* Accounting,
* Relationship with P2P computing,
* Programming utilities, tools and libraries,
* Application-related Issues,
* Data management tools and techniques,
* Simulation and performance modelling,
* Applications,
* Events, news and general announcements.
END CHARTER.
PROCEDURE:
This is a request for discussion, not a call for votes. In this phase
of the process, any potential problems with the proposed newsgroups
should be raised and resolved. The discussion period will continue for
a minimum of 21 days (starting from when the first RFD for this proposal
is posted to news.announce.newgroups), after which a Call For Votes
(CFV) will be posted by a neutral vote taker. Please do not attempt to
vote until this happens.
All discussion of this proposal should be posted to news.groups.
This RFD attempts to comply fully with the Usenet newsgroup creation
guidelines outlined in "How to Create a New Usenet Newsgroup" and "How
to Format and Submit a New Group Proposal". Please refer to these
documents (available in news.announce.newgroups) if you have any
questions about the process.
END PROCEDURE.
DISTRIBUTION: comp.p2p-grid
This RFD has been posted to the following newsgroups:
news.announce.newgroups
news.groups
comp.arch
comp.parallel
comp.parallel.pvm
comp.parallel.mpi
comp.sys.super
comp.client-server
and to the following mailing list:
<gridco...@yahoogroups.com>
END DISTRIBUTION.
Proponent: Rajkumar Buyya <rajk...@csse.monash.edu.au>
Proponent: Mark Baker <mark....@computer.org>
ru.ig...@usask.ca
> REQUEST FOR DISCUSSION (RFD)
> unmoderated group comp.p2p-grid
[snip]
>Newsgroup line:
>comp.p2p-grid Peer-to-Peer / Grid Computing.
I'm wondering about the name. I understand that this is has a
wide scope, yet it is still a subset of peer-to-peer technology.
So part of my brain asks, wouldn't it be more appropriate to
call it comp.p2p.grid (even though there isn't a c.p2p yet)?
Another part of my brain asks if this actually belongs in
something like comp.sys.p2p..., though the rest of my brain
says, don't be an idiot, most comp.sys subgroups are for specific
computer companies or models. Another part asks if this is
another form of parallel processing and if it belongs under
c.parallel? I guess I'm asking, why does p2p-grid deserve
its own 2nd level position, and where does that leave p2p in
general (it's not like that's a new concept)?
>RATIONALE: comp.p2p-grid
Ok, fine. Keep in mind that the most important factor in creating
a new Big-8 news group is having enough readers (not how important
the topic is, or how accessible the group will make the topic). The
fact is, generally, nobody here cares about descriptions such as the
above (but there's nothing wrong with having it). The function of
news.groups is to deal with the technical issues of each proposal.
That is, "the need" you want to show actually must be in how many
and why that many folks want the new group, not the impending spread
of this technology. With that in mind...
>Although over 20 discussion mailing lists operated by individuals or
>institutions exist,
You should try to find out how many subscribers there are and
include that information right here. Try to give some indication
of how much of those lists seem interested in the new group.
>they are generally meant for discussion of
>specific issues (e.g., standards).
Explain why this is a problem in a followup sentence. And perhaps
indicate the degree of the problem.
>Another concern is that mailing
>lists are likely to generate large volume of email for members;
Are you saying they aren't right now? Then from this standpoint
there's no need for a new news group yet, the arguement being
that there isn't enough traffic to matter right now. Otherwise,
your use of "are likely to" is probably bad choice of wording.
Clarify.
>therefore many prospective participants feel discouraged from
>subscribing, do not become members, and do not join important topical
>discussions. Instead, many prefer to participate in discussions as
>and when they want without getting swamped with emails.
Is this your perception, or is this what measured number of
folks have been saying? Be a little more specific about the
source of this information, and perhaps indicate the degree
of the problems, if you can.
>Therefore, we
>believe having a newsgroup dedicated to Grid computing helps overcome
>all these limitations and will encourage discussion and dissemination
>without the need of explicit membership.
Yes, but are there a lot of readers ready to read the group right
now. You are going to need a couple thousand folks supporting
the idea right now in order for you to get the 120 necessary
YES votes (about 1 in 10 actually votes). Only that, and the
factors that contribute to that, are what you should focus on.
>Strategy for publicising the comp.p2p-grid newsgroup:
>The formation of the comp.p2p-grid newsgroup will be publicised through
>the following channels (but not limited to):
[list of channels snipped]
Publicizing during the RFD stage is not a big issue (though you
should have done it before the RFD if you wanted to attract
readers and voters). However, be aware there are restrictions
in what you can post during the vote (i.e. you can only put
out neutral pointers to the fact that there is a vote going
on, you can't ask for support or whatever). Ask the votetaker
when the time comes, or the group creation mentors. After the
vote, go nuts (though that could get some folks mad). :)
>END RATIONALE.
>CHARTER: comp.p2p-grid
This section should have one or two introductory statements
on what this group is about, i.e. what does "p2p" stand for,
and what is the significance of the "grid" qualifier. And
for the sake of consistency, this section should start out with
something like,
"Comp.p2p-grid is an unmoderated newsgroup for discussions
of peer-to-pear (p2p)..."
>Appropriate areas of discussion would include (but not limited to):
^are not
> * Metacomputing, meta-problems,
> * Wide area distributed computing,
Is the above limited to p2p, or is it covered in a group like
c.parallel already? That is, should this be a more specific
statement?
> * Protocols and interfaces for remote access to resources,
> * Local and distributed resource management,
> * Security, authentication, and authorization policies,
> * Resource brokers,
> * Scheduling policies,
> * Computational economy,
> * Quality of Services,
> * Accounting,
You mean chartered accountants are involved, too? :)
> * Relationship with P2P computing,
> * Programming utilities, tools and libraries,
> * Application-related Issues,
> * Data management tools and techniques,
> * Simulation and performance modelling,
> * Applications,
Are some of these kind of obvious or generally applicable, e.g.
"Quality of Services", "Computational economy", "Application-related
issues" (in light of "Applications" 3 lines down) seem that way to me.
A list here could contain a couple obvious ones, but I'd be more
interested in topics that might be more indicative of the topic,
like your "protocols and interfaces" or "resource management".
Otherwise, the list seems to tend towards bits of verbiage. It's
not that big a deal, though.
> * Events, news and general announcements.
Um, what kind of events, news, and general announcments? I think
you might want to add "related to the field of peer-to-peer...",
in case some newbie gets the odd idea that any kind of news or
current events is ok to post. Remember, "without the neeed of
explicit membership" also means the general public can drop a
message into the group (which also means spammers but this being
a technical group might avoid a lot of those).
Which reminds me, are there any topics that would NOT be appropriate,
though they might look appropriate?
Oh yeah, on the announcments, you might want to put a frequency
limit in. I think the usual suggestions are "not more than once
a month" (some times "not more than once in two weeks").
>END CHARTER.
[snip]
>DISTRIBUTION: comp.p2p-grid
>This RFD has been posted to the following newsgroups:
>news.announce.newgroups
>news.groups
>comp.arch
>comp.parallel
>comp.parallel.pvm
>comp.parallel.mpi
>comp.sys.super
>comp.client-server
Is this going to run into the crosspost limit filter problem (i.e. some
ISPs filter out any postings with 5 or more groups)? Or is this so
technical that the ISPs that matter will not affect the situation?
If this is a concern, the list might need to be trimmed and a separate
list be made of groups where a separate "pointer" will be posted
(point them to news.groups for discussion, news.announce.newgroups
for announcements). And in case the subject comes up, it is ok
to post discussion outside of news.groups (if the other groups don't
mind it), but be sure a copy is crossposted to news.groups. At this
point, you want to make sure the technical aspects in any discussion
are can be addressed by those "in the know".
Rajkumar Buyya
ru.ig...@usask.ca wrote:
>
> In news.groups Rajkumar Buyya <rajk...@csse.monash.edu.au> wrote:
> > REQUEST FOR DISCUSSION (RFD)
> > unmoderated group comp.p2p-grid
>
> [snip]
>
> >Newsgroup line:
> >comp.p2p-grid Peer-to-Peer / Grid Computing.
>
> I'm wondering about the name. I understand that this is has a
> wide scope, yet it is still a subset of peer-to-peer technology.
> So part of my brain asks, wouldn't it be more appropriate to
> call it comp.p2p.grid (even though there isn't a c.p2p yet)?
> Another part of my brain asks if this actually belongs in
> something like comp.sys.p2p..., though the rest of my brain
> says, don't be an idiot, most comp.sys subgroups are for specific
> computer companies or models. Another part asks if this is
> another form of parallel processing and if it belongs under
> c.parallel? I guess I'm asking, why does p2p-grid deserve
> its own 2nd level position, and where does that leave p2p in
> general (it's not like that's a new concept)?
In principle both P2P and Grid have many things in common. Initially, I proposed
it as comp.grid and after having discussion with moderators, it is being changed.
The resources in Grid owned by peers who may be colleagues, another organization
freely sharing or a vendor selling/rending services. Both P2P and Grid computing
can be discussed here.
Conceptually/Technicaly, both P2P and Grid are same except that first one (Grid) comes from
people involves in traditional HPC/Supercomputing; and other comes from those with parameter
processing/large data-sets processing applications using PC across the Internet. It will be
interesting to get both worlds in a common place.
This field is a combination of two key things: Parallel and Distributed Computing.
So, just placing in comp.parallel may not be good idea.
good idea.
>
> >they are generally meant for discussion of
> >specific issues (e.g., standards).
>
> Explain why this is a problem in a followup sentence. And perhaps
> indicate the degree of the problem.
>
> >Another concern is that mailing
> >lists are likely to generate large volume of email for members;
>
> Are you saying they aren't right now? Then from this standpoint
> there's no need for a new news group yet, the arguement being
> that there isn't enough traffic to matter right now. Otherwise,
> your use of "are likely to" is probably bad choice of wording.
> Clarify.
They already generate large email traffic. Therefore, I for one
choose to go for Digest mode.
>
> >therefore many prospective participants feel discouraged from
> >subscribing, do not become members, and do not join important topical
> >discussions. Instead, many prefer to participate in discussions as
> >and when they want without getting swamped with emails.
>
> Is this your perception, or is this what measured number of
> folks have been saying? Be a little more specific about the
> source of this information, and perhaps indicate the degree
> of the problems, if you can.
For some of the mailing lists that I had setup, number of colleague
do not want to join for fear of email trafic.
>
> >Therefore, we
> >believe having a newsgroup dedicated to Grid computing helps overcome
> >all these limitations and will encourage discussion and dissemination
> >without the need of explicit membership.
>
> Yes, but are there a lot of readers ready to read the group right
> now. You are going to need a couple thousand folks supporting
> the idea right now in order for you to get the 120 necessary
> YES votes (about 1 in 10 actually votes). Only that, and the
> factors that contribute to that, are what you should focus on.
Ok.
>
> >Strategy for publicising the comp.p2p-grid newsgroup:
>
> >The formation of the comp.p2p-grid newsgroup will be publicised through
> >the following channels (but not limited to):
>
> [list of channels snipped]
>
> Publicizing during the RFD stage is not a big issue (though you
> should have done it before the RFD if you wanted to attract
> readers and voters). However, be aware there are restrictions
> in what you can post during the vote (i.e. you can only put
> out neutral pointers to the fact that there is a vote going
> on, you can't ask for support or whatever). Ask the votetaker
> when the time comes, or the group creation mentors. After the
> vote, go nuts (though that could get some folks mad). :)
Ok.
>
> >END RATIONALE.
>
> >CHARTER: comp.p2p-grid
>
> This section should have one or two introductory statements
> on what this group is about, i.e. what does "p2p" stand for,
> and what is the significance of the "grid" qualifier. And
> for the sake of consistency, this section should start out with
> something like,
> "Comp.p2p-grid is an unmoderated newsgroup for discussions
> of peer-to-pear (p2p)..."
Thanks.
>
> >Appropriate areas of discussion would include (but not limited to):
> ^are not
>
> > * Metacomputing, meta-problems,
> > * Wide area distributed computing,
>
> Is the above limited to p2p, or is it covered in a group like
> c.parallel already? That is, should this be a more specific
> statement?
They are hardly discussed in c.parallel;
>
> > * Protocols and interfaces for remote access to resources,
> > * Local and distributed resource management,
> > * Security, authentication, and authorization policies,
> > * Resource brokers,
> > * Scheduling policies,
> > * Computational economy,
> > * Quality of Services,
> > * Accounting,
>
> You mean chartered accountants are involved, too? :)
No. I mean measuring and accounting of resource consumption so that we
can enable exchange and sharing of resources for public good or profit
in a controlled and regulated manner.
>
> > * Relationship with P2P computing,
> > * Programming utilities, tools and libraries,
> > * Application-related Issues,
> > * Data management tools and techniques,
> > * Simulation and performance modelling,
> > * Applications,
>
> Are some of these kind of obvious or generally applicable, e.g.
> "Quality of Services", "Computational economy", "Application-related
> issues" (in light of "Applications" 3 lines down) seem that way to me.
> A list here could contain a couple obvious ones, but I'd be more
> interested in topics that might be more indicative of the topic,
> like your "protocols and interfaces" or "resource management".
> Otherwise, the list seems to tend towards bits of verbiage. It's
> not that big a deal, though.
>
OK.
> > * Events, news and general announcements.
>
> Um, what kind of events, news, and general announcments? I think
> you might want to add "related to the field of peer-to-peer...",
> in case some newbie gets the odd idea that any kind of news or
> current events is ok to post. Remember, "without the neeed of
> explicit membership" also means the general public can drop a
> message into the group (which also means spammers but this being
> a technical group might avoid a lot of those).
I agree.
>
> Which reminds me, are there any topics that would NOT be appropriate,
> though they might look appropriate?
Grid/P2P technical announcements is fine. We need to avoid sales specific advertisements.
>
> Oh yeah, on the announcments, you might want to put a frequency
> limit in. I think the usual suggestions are "not more than once
> a month" (some times "not more than once in two weeks").
I agree.
You have raised many good points.
Thanks.
Raj
--
Jay Denebeim
Rajkumar Buyya <rajk...@csse.monash.edu.au> wrote:
>Both P2P and Grid computing can be discussed here.
Is Grid Computing still in business? I'm talking about the hardware
company here.
Jay
--
* Jay Denebeim Moderator rec.arts.sf.tv.babylon5.moderated *
* newsgroup submission address: b5...@deepthot.org *
* moderator contact address: b5mod-...@deepthot.org *
* personal contact address: dene...@deepthot.org *
Meg Worley
Ru wrote:
>> That is, "the need" you want to show actually must be in how many
>> and why that many folks want the new group, not the impending spread
>> of this technology.
I'd like to see Rajkumar show how much discussion there
currently is *on Usenet* about grid computing. (Ru, I can't
believe you didn't chant your mantra!) Where is it being
discussed now? How many posts on it per week or day in comp.
parallel or elsewhere? That's what really counts, not people
reading email lists or web forums, and your rationale needs
that badly.
(Particularly since NASA and various other places seem to be
abandoning their work on the Grid, I have real reservations
about demand for such a newsgroup, and especially a second-level
one.)
Rage away,
meg
--
Meg Worley _._ m...@steam.stanford.edu _._ Comparatively Literate
ru.ig...@usask.ca
>Ru wrote:
>>> That is, "the need" you want to show actually must be in how many
>>> and why that many folks want the new group, not the impending spread
>>> of this technology.
>I'd like to see Rajkumar show how much discussion there
>currently is *on Usenet* about grid computing. (Ru, I can't
>believe you didn't chant your mantra!)
Yeah, yeah, yeah. I got sidetracked. If it makes anyone
happy, here it is:
"Quality, usefulness, or merit of a topic is more or less
irrelevant for creating Big-8 newsgroups. Usenet popularity
is the primary consideration."
>Where is it being
>discussed now? How many posts on it per week or day in comp.
>parallel or elsewhere? That's what really counts, not people
>reading email lists or web forums, and your rationale needs
>that badly.
I basically got hung up on telling the proponent that his mailing
list information is only of use if the readers are willing to
go to the newsgroup. I forgot to deal with the usenet
readership side of things. Regardless, Meg has a point that
should be addressed (in the next version of the RFD). And
let me reiterate my earlier point, when you go looking at
the mailing list readership, you need to find out approximately
how many seem to be willing to read the new group. Otherwise,
Meg's point should be taken: our experience is that the mere
presence of mailing lists and web forums means nearly ZERO for
creating a Big-8 newsgroup, and your only usable fallback is
usenet volume.
That reminds me, proponents, a lot of my comments and questions
in my earlier posting is intended for you to add your responses
to the next version of the RFD, not just for you to respond in
this forum.
ru
piranha
>
> I'm wondering about the name. I understand that this is has a
> wide scope, yet it is still a subset of peer-to-peer technology.
> So part of my brain asks, wouldn't it be more appropriate to
> call it comp.p2p.grid (even though there isn't a c.p2p yet)?
> Another part of my brain asks if this actually belongs in
> something like comp.sys.p2p..., though the rest of my brain
> says, don't be an idiot, most comp.sys subgroups are for specific
> computer companies or models. Another part asks if this is
> another form of parallel processing and if it belongs under
> c.parallel? I guess I'm asking, why does p2p-grid deserve
> its own 2nd level position, and where does that leave p2p in
> general (it's not like that's a new concept)?
thank you, ru. i no longer feel alone in my vascillations. :-)
--
-piranha
Shrisha Rao
Rajkumar Buyya <rajk...@csse.monash.edu.au> wrote:
> REQUEST FOR DISCUSSION (RFD)
> unmoderated group comp.p2p-grid
>
>This is a formal Request For Discussion (RFD) for the creation of
>world-wide unmoderated Usenet newsgroup comp.p2p-grid. This is
>not a Call for Votes (CFV); you cannot vote at this time. Procedural
>details are below.
>
>Newsgroup line:
>comp.p2p-grid Peer-to-Peer / Grid Computing.
>
>
>RATIONALE: comp.p2p-grid
>
>Grid Computing is a new paradigm that enables sharing of geographically
>distributed resources such as computers, databases, and scientific
>instruments in a secure and consistent manner without any central control
>for solving large-scale problems in science, engineering, and commerce.
The description is not meaningful to me. As stated, it better fits
distributed computing, which is hardly a new paradigm, though it has
gained popularity of late.
The taxonomy of the comp.* groups leaves much to be desired, but it
may be better to have a comp.distributed.misc and then various
specific variants, including embedded computing, peer-to-peer
computing, etc., underneath. Having everything under comp.* is a
mistake that need not be made worse by repetition.
The name should also probably be ended with a *.misc, and use either
`p2p' or `grid' only. Having both doesn't seem natural or necessary.
>Strategy for publicising the comp.p2p-grid newsgroup:
>
>The formation of the comp.p2p-grid newsgroup will be publicised through
>the following channels (but not limited to):
>
> * IEEE DS Online,
> * Global Grid Forum,
> * P2P WG,
> * Grid Infoware,
> * IEEE/ACM conferences:
Can you name some IEEE/ACM conferences that might be appropriate? I
don't think either ACM-PODC (Principles of Distributed Computing --
see http://www.podc.org) or similar ones would be.
I don't think the distributed computing community is on Usenet to any
significant extent, though some individuals who belong to it are.
There are some mailing lists (announcement rather than discussion) and
that seems to be the preferred mode of communication.
Regards,
Shrisha Rao
Rajkumar Buyya
Jay Denebeim wrote:
>
> In article <3BC41557...@csse.monash.edu.au>,
> Rajkumar Buyya <rajk...@csse.monash.edu.au> wrote:
>
> >Both P2P and Grid computing can be discussed here.
>
> Is Grid Computing still in business? I'm talking about the hardware
> company here.
which hardware company ? There are many software companies are jumping
into this space including Sun, Microsoft and many startups.
BTW, Microsoft'w Gordon Bell think ".NET is Grid"--see his remarks
in news article:
http://it.mycareer.com.au/cgi-bin/common/printArticle.pl?path=/news/2001/10/09/FFXMLM06JSC.html
OR: http://it.mycareer.com.au/news/2001/10/09/FFXMLM06JSC.html
Gordon comment reads as:
---------
"(Microsoft's next generation architecture) .NET is grid," says Microsoft researcher Gordon
Bell, who has worked on the grid problem with colleague Jim Gray. "Let's face it, it's a
commercial version of grid, so to me, the grid is quite interesting because it's the ultimate
in distributed computing."
---------
cheers
Raj
>
> Jay
>
> --
> * Jay Denebeim Moderator rec.arts.sf.tv.babylon5.moderated *
> * newsgroup submission address: b5...@deepthot.org *
> * moderator contact address: b5mod-...@deepthot.org *
> * personal contact address: dene...@deepthot.org *
--
Best regards,
Raj
------------------------------------------------------------------------
Rajkumar Buyya
School of Computer Science and Software Engineering
Monash University, C5.41, Caulfield Campus
Melbourne, VIC 3145, Australia
Phone: +61-3-9903 1969 (office); +61-3-9571 3629 (home)
Fax: +61-3-9903 2863; eFax: +1-801-720-9272
Email: rajk...@buyya.com | rajk...@csse.monash.edu.au
URL: http://www.buyya.com | http://www.csse.monash.edu.au/~rajkumar
------------------------------------------------------------------------
Rajkumar Buyya
Meg Worley wrote:
>
> Ru wrote:
> >> That is, "the need" you want to show actually must be in how many
> >> and why that many folks want the new group, not the impending spread
> >> of this technology.
>
> I'd like to see Rajkumar show how much discussion there
> currently is *on Usenet* about grid computing. (Ru, I can't
> believe you didn't chant your mantra!) Where is it being
> discussed now? How many posts on it per week or day in comp.
> parallel or elsewhere? That's what really counts, not people
> reading email lists or web forums, and your rationale needs
> that badly.
Grid/P2P is not much discussed on Usnet group, apart from annoucements
on comp.parallel and related groups. However, discussion happens in major
scale on mailing lists, eg., GGF and decentralization as yahoo group.
> (Particularly since NASA and various other places seem to be
> abandoning their work on the Grid, I have real reservations
> about demand for such a newsgroup, and especially a second-level
> one.)
I am not aware of NASA abandoning their work on the Grid. Any pointers
to information on abandoning Grid words ?
More and more projects are comming up with backing from many govt.
and industry research investments.
Thanks
Raj
>
> Rage away,
>
> meg
>
> --
>
> Meg Worley _._ m...@steam.stanford.edu _._ Comparatively Literate
--
ru.ig...@usask.ca
>ru.ig...@usask.ca wrote:
>>
>> In news.groups Rajkumar Buyya <rajk...@csse.monash.edu.au> wrote:
>> > REQUEST FOR DISCUSSION (RFD)
>> > unmoderated group comp.p2p-grid
>>
[snip]
>> I guess I'm asking, why does p2p-grid deserve
>> its own 2nd level position,
>In principle both P2P and Grid have many things in common. Initially, I proposed
>it as comp.grid and after having discussion with moderators, it is being changed.
>The resources in Grid owned by peers who may be colleagues, another organization
>freely sharing or a vendor selling/rending services. Both P2P and Grid computing
>can be discussed here.
Ok, let me concentrate on one aspect of the name issue, that of the
name level it is being proposed at. It is on the second tier. That
would make it a peer of such component or fundamental areas as:
- OS
- sources
- software
- sys
- lang
- graphics
- dcom
- protocols
and some general application areas such as:
- windows
- unix
- cad
- databases
- ai
- emulators
- publish
- org
Does p2p-grid fit into either category? The way it has been described
I question a placement in the first category as it seems more of an
implementation than a new component unto itself. Yet it's being an
implementation almost of puts it on the second list. Almost, because
p2p-grid isn't really an application concept like them either, it's
infrastructure. Also, one has to keep in mind that some of those
groups were created before there was much thought put into systemization.
Also, regardless of those considerations, a second level group
should be expected to split in the future (if not right away).
When I think back on some of the topics in the charter, I can
see what some of though subgroups might be, such as protocols,
security, and software. And when one looks at other comp.*
groups, many terminate at that (3rd) level. Of course, that
still doesn't exclude the idea of pushing p2p-grid to 3rd level
with terminating groups at 4th level. And indeed, it looks
to me that the kinds of forseeable splits are the kinds that
we currently see at the 4th level.
Another thing that's bothering me is that I think there was a
statement that p2p and grid overlapped and that one is not a
subset of the other. That kind of makes the name look like
a combined general P2P and general grid discussion group. Indeed,
I believe there was a statement along the lines that p2p discussion
would be acceptable. Then, is there a more general term that
covers p2p and grid computing, that doesn't look as half-baked
as "p2p-grid"? If so, it's possible THAT name would be appropriate
for a second level group.
But the one thing that I keep bouncing back (from the rubber walls)
to is that in principle, this LOOKS like another "sys". That is,
one is glomming a bunch of spatially and OS disparate boxes together
to compute, like the network topography is a superset of motherboard
topography. Sure it's a new paradigm, but to me it looks like a
new paradigm in the layout of what one could call a computer. So
I end up asking, "really, why couldn't this go under comp.sys?"
I think those thoughts are what have been itching away at the
back of my mind ever since I first skimmed the proposal. At
least, I do know that I find the name somehow irritating.
ru
Rajkumar Buyya
> In article <10026657...@isc.org>,
> Rajkumar Buyya <rajk...@csse.monash.edu.au> wrote:
> > REQUEST FOR DISCUSSION (RFD)
> > unmoderated group comp.p2p-grid
> >
> >This is a formal Request For Discussion (RFD) for the creation of
> >world-wide unmoderated Usenet newsgroup comp.p2p-grid. This is
> >not a Call for Votes (CFV); you cannot vote at this time. Procedural
> >details are below.
> >
> >Newsgroup line:
> >comp.p2p-grid Peer-to-Peer / Grid Computing.
> >
> >
> >RATIONALE: comp.p2p-grid
> >
> >Grid Computing is a new paradigm that enables sharing of geographically
> >distributed resources such as computers, databases, and scientific
> >instruments in a secure and consistent manner without any central control
> >for solving large-scale problems in science, engineering, and commerce.
>
> The description is not meaningful to me. As stated, it better fits
> distributed computing, which is hardly a new paradigm, though it has
> gained popularity of late.
>
> The taxonomy of the comp.* groups leaves much to be desired, but it
> may be better to have a comp.distributed.misc and then various
> specific variants, including embedded computing, peer-to-peer
> computing, etc., underneath.
Grid/P2P is certainly a type of "Parallel and Distributed Computing" and has flavors of both. I
believe Grid world is mostly driven by those with major focus in Parallel Computing, but on
Distributed Systems (distributed HPC machines or low-end machines like PCs).
This is the only confusion, otherwise, it is fine to place it in comp.distributed
as comp.distributed.grid and comp.distributed.p2p; But then, it looks like we do not
have a newsgroup on "comp.distributed". I am I right ?
If 2nd level group creation is not required, it is fine to place in:
comp.parallel.grid or comp.parallel.p2p
It can create a confusion about whether it is parallel or distributed computing. In such cases,
it might be worth placing it in: comp.research.grid and comp.research.p2p since many efforts in
this area still in research and building phase.
> Having everything under comp.* is a
> mistake that need not be made worse by repetition.
>
> The name should also probably be ended with a *.misc, and use either
> `p2p' or `grid' only. Having both doesn't seem natural or necessary.
>
> >Strategy for publicising the comp.p2p-grid newsgroup:
> >
> >The formation of the comp.p2p-grid newsgroup will be publicised through
> >the following channels (but not limited to):
> >
> > * IEEE DS Online,
> > * Global Grid Forum,
> > * P2P WG,
> > * Grid Infoware,
> > * IEEE/ACM conferences:
>
> Can you name some IEEE/ACM conferences that might be appropriate?
Check out: CCGrid conference: http://www.ccgrid.org
http://www.gridcomputing.org
> don't think either ACM-PODC (Principles of Distributed Computing --
> see http://www.podc.org) or similar ones would be.
>
> I don't think the distributed computing community is on Usenet to any
> significant extent, though some individuals who belong to it are.
> There are some mailing lists (announcement rather than discussion) and
> that seems to be the preferred mode of communication.
I agree.
Looks like: comp.client-server group is active.
Regards
Raj
>
> Regards,
>
> Shrisha Rao
>
> spammer...@yahoogroups.com
> --
>
> http://www.dvaita.org
> http://www.dvaita.net
--
ru.ig...@usask.ca
>Meg Worley wrote:
>>
>> Ru wrote:
>> >> That is, "the need" you want to show actually must be in how many
>> >> and why that many folks want the new group, not the impending spread
>> >> of this technology.
>>
>> I'd like to see Rajkumar show how much discussion there
>> currently is *on Usenet* about grid computing. (Ru, I can't
>> believe you didn't chant your mantra!) Where is it being
>> discussed now? How many posts on it per week or day in comp.
>> parallel or elsewhere? That's what really counts, not people
>> reading email lists or web forums, and your rationale needs
>> that badly.
>Grid/P2P is not much discussed on Usnet group, apart from annoucements
>on comp.parallel and related groups.
In general, that's a bit odd. One would think that if a comp
topic has gotten hot enough, discussions would have bled into some
of the related groups (which then would force a call for a new
group because the on-topic stuff starts to suffer). That is,
if the mailing list readers are currently willing to use usenet,
why aren't they? There ought to be one comp.* group that is
close enough to the topic that a few announcements in the
mailing lists would have moved a bunch of discussion over.
I mean, if there are mailing lists for one or two highly
specialized aspects of this topic, wouldn't they have a related
(possibly more general) newsgroup? You might want to address
this in the next version of the RFD.
>However, discussion happens in major
>scale on mailing lists, eg., GGF and decentralization as yahoo group.
Define "major scale" quantitatively. Our experience here has
been that someone that uses that kind of language either
underestimates what is considered large, or is exaggerating.
Even then, just knowing the size and traffic volume of the mailing
lists is of doubtful use.
One has to keep in mind that sometimes a majority of users of
mailing lists actually want to stay with mailing lists
because there is better focus in the discussions, keeps out spam,
etc, even if it means others are scared off (e.g. if you are
scared off, you aren't serious enough for what needs to be
discussed).
Bottom line: don't count on the mailing lists (unless they tell
you to)
ru
ru.ig...@usask.ca
>This is the only confusion, otherwise, it is fine to place it in comp.distributed
>as comp.distributed.grid and comp.distributed.p2p; But then, it looks like we do not
>have a newsgroup on "comp.distributed". I am I right ?
I personally wouldn't object to that situation, but if you are worried
about it, just propose the general group with the other two. I think
your concern about "parallel vs distributed" is the bigger problem.
>If 2nd level group creation is not required, it is fine to place in:
> comp.parallel.grid or comp.parallel.p2p
>It can create a confusion about whether it is parallel or distributed computing. In such cases,
>it might be worth placing it in: comp.research.grid and comp.research.p2p since many efforts in
>this area still in research and building phase.
I think it best not to go this route of interim placement. It's a
pain in the rear moving to a new (and more appropriate) name afterwards.
If the problem is p2p and grid have enough commonality that there
needs to be a union group, then think about a concept that contains
both and try that for the name of a general or nexus group for the
two. It seems "parallel" and "distributed" are equally different
and similar as "p2p" and "grid" are, so avoid those, or use those to
guide you to an umbrella name. That is, what do "parallel
computing", "distributed computing", "p2p", and "grid computing"
have in common? "Computing"... damn *smack*. Well, they are
multiprocessing, aren't they?
[attribution lost]
>>
>> The name should also probably be ended with a *.misc
This I don't agree with in general. I've always thought the
idea of a .misc tag sucked, at least thematically.
ru
Shrisha Rao
>That is, what do "parallel
>computing", "distributed computing", "p2p", and "grid computing"
>have in common?
I can't speak for the latter two, but broadly speaking, one could say
that "parallel computing" is a phrase more often used in case of
multiprocessor architectures where memory is shared, while
"distributed computing" is more often applied in case of
message-passing. This is not rigorous, and a lot of theoretical work
(such as in PODC) also references shared memory.
In fact, the theoretical research on distributed computing has little
to do with practice, but that's a whole other issue.
>[attribution lost]
>>>
>>> The name should also probably be ended with a *.misc
That was me.
>This I don't agree with in general. I've always thought the
>idea of a .misc tag sucked, at least thematically.
Well, no one's questioning your right to your opinion.
Regards,
Shrisha Rao
>ru
Phil Gustafson
>
>Jay Denebeim wrote:
>>
>> Rajkumar Buyya <rajk...@csse.monash.edu.au> wrote:
>>
>> >Both P2P and Grid computing can be discussed here.
>>
>> Is Grid Computing still in business? I'm talking about the hardware
>> company here.
>
>which hardware company ? There are many software companies are jumping
>into this space including Sun, Microsoft and many startups.
>
years ago. These were very studly magnesium gadgets with bubble memory,
sort of Mercedes computers competing with Adam Osborne's Pintos.
There's still a vestige of the company at (duh) www.grid.com; they
apparently mostly service military GRiDs.
These certainly aren't what the proposed group is about, and there might
some small confusion if it were named comp.sys.grid.
Phil
--
Phil Gustafson <ph...@panix.com>
There is a tarantula in the song, of course. I remember when you
could ask your grocer to save one for you. -- Casady
Jay Denebeim
Rajkumar Buyya <rajk...@csse.monash.edu.au> wrote:
>which hardware company ?
Grid. That's the name of the company, they make/made stealth laptops.
Spooks and the military used them because they were designed for 0
emissions and were therefore supposed to be undetectable.
>There are many software companies are jumping
>into this space including Sun, Microsoft and many startups.
Not at all what I'm talking about.
>BTW, Microsoft'w Gordon Bell think ".NET is Grid"--see his remarks
>in news article:
Okay, so what you're saying here is the evil empire wants to use grid
to steal the net from us? I really suggest you don't go there, there
are alot of people who are really pissed at microsoft over that thing
and your group is likely to get painted with that brush if you invoke
the meme.
Rajkumar Buyya
P2P/Grid Computing is concept. It is worth placing it in the first catagority.
Some of the key application areas of the way P2P/Grid is practiced today are:
* Distributed Supercomputing:
for executing parallel applications on distributed supercomputers.
* High Throughput Computing:
Large scale simulation/chip design & parameter studies.
* Content Sharing (free or paid)
Sharing digital contents among peers (e.g., Napster)
* Remote software access/renting services:
Application service provides (ASPs) & Web services.
* Data-intensive computing:
Virtual Drug Design, Particle Physics, Stock Prediction...
* On-demand, realtime computing:
Medical instrumentation & Mission Critical.
*Collaborative Computing:
Collaborative design, Data exploration, education.
* Service Oriented Computing (SOC):
Computing as Utility: New paradigm and new industries.
> Also, regardless of those considerations, a second level group
> should be expected to split in the future (if not right away).
Yes, it is possible to split into many levels like:
comp.p2p-grid.scheduling
comp.p2p-grid.resource-management
comp.p2p-grid.directory
comp.p2p-grid.security
comp.p2p-grid.apps (applications)
we can have many application catagory as well.
comp.p2p-grid.data (like data grid)
comp.p2p-grid.protocols
comp.p2p-grid.qos (quality of services)
> When I think back on some of the topics in the charter, I can
> see what some of though subgroups might be, such as protocols,
> security, and software. And when one looks at other comp.*
> groups, many terminate at that (3rd) level. Of course, that
> still doesn't exclude the idea of pushing p2p-grid to 3rd level
> with terminating groups at 4th level. And indeed, it looks
> to me that the kinds of forseeable splits are the kinds that
> we currently see at the 4th level.
>
> Another thing that's bothering me is that I think there was a
> statement that p2p and grid overlapped and that one is not a
> subset of the other.
I agree. Conceptually both P2P and Grid are same. Hence, I personally
don't mind calling either! What differs is the way people choose to
implement them. Also, origin of these two concepts. Currently so called
Grid computing is following horizontal integration of technologies. Whereas
so P2P community is following vertical integration and who system is application
specific (e.g., SETI@Home), but this is rapidly changece. Another difference is
Grid world is focus on distributed high end machines (like clusters, supercomputers)
whereas P2P world is focus more on low-end devices like PCs.
> That kind of makes the name look like
> a combined general P2P and general grid discussion group. Indeed,
> I believe there was a statement along the lines that p2p discussion
> would be acceptable. Then, is there a more general term that
> covers p2p and grid computing, that doesn't look as half-baked
> as "p2p-grid"? If so, it's possible THAT name would be appropriate
> for a second level group.
>
> But the one thing that I keep bouncing back (from the rubber walls)
> to is that in principle, this LOOKS like another "sys". That is,
> one is glomming a bunch of spatially and OS disparate boxes together
> to compute, like the network topography is a superset of motherboard
> topography. Sure it's a new paradigm, but to me it looks like a
> new paradigm in the layout of what one could call a computer. So
> I end up asking, "really, why couldn't this go under comp.sys?"
possible. like comp.sys.super
>
> I think those thoughts are what have been itching away at the
> back of my mind ever since I first skimmed the proposal. At
> least, I do know that I find the name somehow irritating.
I guess, it will be good to have just one name either p2p or grid;
Thanks
Raj
>
> ru
Jay Denebeim
Rajkumar Buyya <rajk...@csse.monash.edu.au> wrote:
>P2P/Grid Computing is concept. It is worth placing it in the first catagority.
>
>Some of the key application areas of the way P2P/Grid is practiced today are:
>
>* Distributed Supercomputing:
> for executing parallel applications on distributed supercomputers.
>* High Throughput Computing:
> Large scale simulation/chip design & parameter studies.
>* Content Sharing (free or paid)
> Sharing digital contents among peers (e.g., Napster)
>* Remote software access/renting services:
> Application service provides (ASPs) & Web services.
>* Data-intensive computing:
> Virtual Drug Design, Particle Physics, Stock Prediction...
>* On-demand, realtime computing:
> Medical instrumentation & Mission Critical.
>*Collaborative Computing:
> Collaborative design, Data exploration, education.
>* Service Oriented Computing (SOC):
> Computing as Utility: New paradigm and new industries.
In what way does this differ from distributed computing? It sounds
identical to me. Or is this just new names for an extremely old
concept?
David B Terrell
> Grid/P2P is not much discussed on Usnet group, apart from annoucements
> on comp.parallel and related groups. However, discussion happens in major
> scale on mailing lists, eg., GGF and decentralization as yahoo group.
If comp.parallel exists, would comp.parallel.grid be appropriate?
It would seem to fit in well with the other subgroups. I don't
think that comp.p2p-grid is appropriate for several already-mentioned
reasons, and I'd vote no on this proposal for that reason alone.
--
David Terrell | "Instead of plodding through the equivalent of
Prime Minister, NebCorp | literary Xanax, the pregeeks go for sci-fi and
d...@meat.net | fantasy: LSD in book form." - Benjy Feen,
http://wwn.nebcorp.com | http://www.monkeybagel.com/ "Origins of Sysadmins"
ru.ig...@usask.ca
>P2P/Grid Computing is concept.
So are a lot of third level groups. In some sense, this argues
that it is NOT worth placement in the first category.
>It is worth placing it in the first catagority.
>Some of the key application areas of the way P2P/Grid is practiced today are:
>* Distributed Supercomputing:
> for executing parallel applications on distributed supercomputers.
>* High Throughput Computing:
> Large scale simulation/chip design & parameter studies.
>* Content Sharing (free or paid)
> Sharing digital contents among peers (e.g., Napster)
>* Remote software access/renting services:
> Application service provides (ASPs) & Web services.
>* Data-intensive computing:
> Virtual Drug Design, Particle Physics, Stock Prediction...
>* On-demand, realtime computing:
> Medical instrumentation & Mission Critical.
>*Collaborative Computing:
> Collaborative design, Data exploration, education.
>* Service Oriented Computing (SOC):
> Computing as Utility: New paradigm and new industries.
Those are why I don't see this as a second level group in
my first category. Many of the first category have more
to do with some of the basic components in computing.
The proposed topic isn't really a component unto itself.
The other reason I don't see this as a first category
group is that p2p and grid computing deal with the use
of single processors at some point in the structure, and
that is an implementation (concept), not a foundation (concept).
The very fact that p2p and grid can be categorized under
something like distributed or parallel which in turn can
be categorized under something else (say for arguement
sake, multiprocessing) indicates to me that there is a
level between comp and p2p/grid. Actually, when I
think about it, the reason I would not balk nearly as
much if comp.distributed or comp.parallel were proposed
is because THOSE concepts are not dependent on implementation.
>> Also, regardless of those considerations, a second level group
>> should be expected to split in the future (if not right away).
>Yes, it is possible to split into many levels like:
>
>comp.p2p-grid.scheduling
>comp.p2p-grid.resource-management
>comp.p2p-grid.directory
>comp.p2p-grid.security
>comp.p2p-grid.apps (applications)
> we can have many application catagory as well.
>comp.p2p-grid.data (like data grid)
>comp.p2p-grid.protocols
>comp.p2p-grid.qos (quality of services)
Which doesn't argue against the following.
>> And indeed, it looks
>> to me that the kinds of forseeable splits are the kinds that
>> we currently see at the 4th level.
[snip]
>I guess, it will be good to have just one name either p2p or grid;
Yes, one name, but based on what has been posted so far, neither
would be appropriate. You need an umbrella name that can
accomodate both.
ru
Meg Worley
Raj writes:
>Grid/P2P is not much discussed on Usnet group, apart from annoucements
>on comp.parallel and related groups.
That makes it an unlikely candidate for a successful usenet group
for the moment. The Field of Dreams argument ("If you build it,
they will come") has been shown time and time again to be false
with respect to newsgroups.
>However, discussion happens in major
>scale on mailing lists, eg., GGF and decentralization as yahoo group.
As Ru has pointed out, people on mailing lists often don't want to
move. Even when they do, Usenet isn't necessarily where they want
to move *to* -- it's an acquired taste and not for everyone. That's
why the rationale should address current traffic on Usenet.
I wrote:
>> (Particularly since NASA and various other places seem to be
>> abandoning their work on the Grid, I have real reservations
>> about demand for such a newsgroup, and especially a second-level
>> one.)
Raj writes:
>I am not aware of NASA abandoning their work on the Grid. Any pointers
>to information on abandoning Grid words ?
Nah -- I just know that most of the Grid folks here at NASA-Ames
have been given their walking papers, and supposedly the Ames and
Glenn programs have been massively slashed too.
Dave DiNucci
I personally regard the p2p-grid topic as cutting across almost all of
the topics above--i.e. the unique issues that come into consideration
when trying to exploit and combine these fundamental or application
areas in a shared heterogeneous distributed environment. That is, you
could consider a "p2p-grid"-like sub-level in almost every one of them,
scattering these discussions all over, or you could create a new second
(or third?) level topic and provide more focus, which is the whole
idea. For this reason, I'd have a hard time saying whether this new
topic belonged more in the first or second category above.
> ... Actually, when I
> think about it, the reason I would not balk nearly as
> much if comp.distributed or comp.parallel were proposed
> is because THOSE concepts are not dependent on implementation.
I agree that the second level (if there is one) should be based on a
concept rather than implementation, explained below.
> >> Also, regardless of those considerations, a second level group
> >> should be expected to split in the future (if not right away).
>
> >Yes, it is possible to split into many levels like:
> >
> >comp.p2p-grid.scheduling
> >comp.p2p-grid.resource-management
> >comp.p2p-grid.directory
> >comp.p2p-grid.security
> >comp.p2p-grid.apps (applications)
> > we can have many application catagory as well.
> >comp.p2p-grid.data (like data grid)
> >comp.p2p-grid.protocols
> >comp.p2p-grid.qos (quality of services)
>
> Which doesn't argue against the following.
>
> >> And indeed, it looks
> >> to me that the kinds of forseeable splits are the kinds that
> >> we currently see at the 4th level.
>
> [snip]
>
> >I guess, it will be good to have just one name either p2p or grid;
>
> Yes, one name, but based on what has been posted so far, neither
> would be appropriate. You need an umbrella name that can
> accomodate both.
FWIW, I have been using the term "distributed [computational] resource
collectives" or "peer-to-grid" to encompass those features common to
both p2p and grids--i.e. (here's the concept) the synergistic
distributed sharing of compute, information, human, and
instrument/peripheral resources. The term "metacomputing" was coined
some time ago by Smarr to mean something very similar. Note that
parallel or distributed computing per se is only one part of the focus
here. The p2p community sometimes uses the term "decentralization" to
describe their focus, but in some cases that can be more of an obstacle
to be overcome than a goal of this approach.
Although there are plenty of common features of p2p and grids (that
should probably be addressed in one group or set of groups), there are
also different focii that are specific to each. For example, grids are
usually considered as shared utilities (like electrical grid),
relatively ubiquitous, pervasive and stable, while p2p is often
considered an on-demand temporary relationship between resources to
satisfy a demand, after which they disband. Grids often aim to provide
unprecedented capabilities to facilitate novel problem solving
approaches, regardless of cost, where p2p is often a cost saving or time
saving efficiency tool for everyday problems. Grid customers are
(currently) often government labs and research institutions, while p2p
is often targeted at smaller businesses and individual consumers. Grid
researchers worry about effective use of high-speed compute and network
resources, while p2p researchers worry about intellectual property
considerations, business models, and connectivity through firewalls and
network address translators (IP masquerading). (Of course, there are
exceptions, hybrids, and middle ground.)
Taking this all into account, I would favor a second-level group under
comp, named something like "collectives" (my favorite) or
"metacomputing" or "distributed" or "peer2grid" or "decentralized", and
assume there will be splits at the third level like those suggested
above which cover many of the common issues faced by both p2p and grids,
as well as other third level splits like "p2p" and "grid" to explicitly
address topics unique to each of these kinds of collectives.
-Dave
-----------------------------------------------------------------
David C. DiNucci http://www.elepar.com Grid & P2P Prog. Tools
Dave DiNucci
>
> Raj writes:
> >Grid/P2P is not much discussed on Usnet group, apart from annoucements
> >on comp.parallel and related groups.
>
> That makes it an unlikely candidate for a successful usenet group
> for the moment. The Field of Dreams argument ("If you build it,
> they will come") has been shown time and time again to be false
> with respect to newsgroups.
I, for one, have occasionally attempted to carry on discussions about
grid topics in a variety of venues, and have found them all lacking. I
believe that an unmoderated Usenet group would satisfy my needs, and I
believe that others are likely to hold this view as well, but I don't
expect the vast majority of such people to follow news.groups. I do
expect that they will vote positively for a new group on the topic.
> >However, discussion happens in major
> >scale on mailing lists, eg., GGF and decentralization as yahoo group.
>
> As Ru has pointed out, people on mailing lists often don't want to
> move. Even when they do, Usenet isn't necessarily where they want
> to move *to* -- it's an acquired taste and not for everyone. That's
> why the rationale should address current traffic on Usenet.
There is no good place on the Usenet to carry on these discussions. The
most success I have had in existing groups was on (I believe)
comp.sys.super, but for the most part, the people who would be
interested in grids don't read it, so it became more an educational
thread rather than a productive one. The same problems apply to
comp.arch, with the added complication that grid discussions get lost in
the vast traffic there. Comp.parallel has been useless for discussions
in recent months (years?), due to some sort of moderation logjam, and
although that may be in the process of being cleared as we speak
(according to Eugene Miya), parallel computing is just one of many grid
issues.
The GGF mailing lists would seem to be a good location for these
discussions, but are not in practice. For example, I have had at least
three discussions on two different lists terminated by GGF group leads.
In one case, the lead said the problem was that my posts were too long,
calling them "spam" (though I posted just one or so per day, and on
topic). I surmise that the real problem, in all cases, was that the
approaches I was discussing weren't obviously compatible with those that
the GGF was currently exploring/implementing. In other words, the GGF
has some specific practical goals to create products (software and/or
white papers), and they apparently and understandably feel that
providing a truly open forum can blur their focus. In fact, even most
discussions of specific GGF topics these days take place outside of the
lists, in private emails, conference calls, and/or white paper drafts.
A Usenet group, on the other hand, would be a truly open forum to
explore and illuminate approaches that are outside of the GGF's focus.
At the first P2P Working Group meeting, there was a brief discussion of
setting up a Usenet group. If I recall, the decentralization Yahoo list
was initiated as sort of a path of least resistance--i.e. they wanted it
fast. (In fact, there other failed attempts, such as
http://pub48.ezboard.com/bpeertopeerworkinggroup) I don't think it had
anything to do with a distaste for Usenet. I believe that a usenet
group will succeed even if the majority of that group doesn't migrate
over.
> I wrote:
> >> (Particularly since NASA and various other places seem to be
> >> abandoning their work on the Grid, I have real reservations
> >> about demand for such a newsgroup, and especially a second-level
> >> one.)
>
> Raj writes:
> >I am not aware of NASA abandoning their work on the Grid. Any pointers
> >to information on abandoning Grid words ?
>
> Nah -- I just know that most of the Grid folks here at NASA-Ames
> have been given their walking papers, and supposedly the Ames and
> Glenn programs have been massively slashed too.
>
> Rage away,
The GGF just met this week in Italy, in spite of all the scares, etc.,
and NASA has announced a big Info Power Grid meeting in December. NSF
has just initiated the Distributed Terascale Facility ("Teragrid") which
involves $53Million, IBM, Intel, Qwest, Oracle, and others.
http://www.nsf.gov/od/lpa/news/press/01/pr0167.htm
IBM recently announced its intention to spend a billion dollars on grids
over the next few years.
http://www.nytimes.com/2001/08/02/technology/02BLUE.html
I don't know who you're referring to at NASA (I'm now sufficiently out
of the loop, and I'm guessing that "most of the Grid folks" is an
exaggeration), but some NASA IPG personnel had temporary assignments in
the first place. Even if NASA IPG itself was flopping as you suggest
(and I would be amazed if it is), it could be blamed on a variety of
factors, not the least of which is the lack of an open venue for
discussion of grid technologies (which some of us have only recently
realized did not exist in the GGF). If you are suggesting that a Usenet
group will go unused because worldwide interest in distributed
computational resource collectives will somehow suddenly evaporate, I
can't fathom it.
Aahz Maruch
>
>Yes, one name, but based on what has been posted so far, neither
>would be appropriate. You need an umbrella name that can
>accomodate both.
comp.distributed gets my vote
--
--- Aahz <*> (Copyright 2001 by aa...@pobox.com)
Hugs and backrubs -- I break Rule 6 http://www.rahul.net/aahz/
Androgynous poly kinky vanilla queer het Pythonista
We must not let the evil of a few trample the freedoms of the many.
Rajkumar Buyya
It is nice to see many important issues and concerns which I was unable
to raise them with other discussion groups. Many of Usnet leaders
feel that 2nd level need to be concepts focus rather that implementation
and also, I noticed that few other colleagues in discussion seem to support
comp.distributed, it might be worth exploring it further. Something like:
comp.distributed.grid
comp.distributed.p2p
We can have many sub level within p2p / grid depending on their relevance.
All common things can be outside as they are need for both P2P and grid world.
I am keen on knowing opinion of other members on this ?
Thanks
Raj
--
Rajkumar Buyya
Aahz Maruch wrote:
>
> In article <9q4s81$phc$1...@tribune.usask.ca>, <ru.ig...@usask.ca> wrote:
> >
> >Yes, one name, but based on what has been posted so far, neither
> >would be appropriate. You need an umbrella name that can
> >accomodate both.
>
> comp.distributed gets my vote
Given so many arguments, I also favour comp.distributed, but two other levels like:
comp.distributed.grid
comp.distributed.p2p
Within then, we can have sublevels as need arises.
Now the questions is: we first need to have: comp.distributed created, or submit
a proposal for comp.distributed.grid, which automatically helps in creating 2nd level
item.
Another hurdle for comp.distributed.grid; is .grid also has parallel computing
concept and in fact many people who have now into grid computing are actually
from Parallel Computing. At this moment, it is like:
Parallel Computing on Wide Area Distributed Resources (Computers: PCs, WSs, clusters, vector
supercomputers).
Thanks
Raj
Jonathan Grobe
>
>Given so many arguments, I also favour comp.distributed, but two other levels
like:
>
>comp.distributed.grid
>comp.distributed.p2p
>
>Within then, we can have sublevels as need arises.
>
>Now the questions is: we first need to have: comp.distributed created, or submit
>a proposal for comp.distributed.grid, which automatically helps in creating 2nd level
>item.
The newsgroup comp.distributed would only be created if you propose
a group with that name. I suggest you just submit a second RFD with
your proposed groups comp.distributed.grid and comp.distributed.p2p
and the more general group comp.distributed.misc (misc is the preferred
name for the general group).
I do seriously wonder if the groups will get any traffic. So far almost
all of the discussion has been from news.groups regulars. Several
months ago I created a group on p2p (alt.internet.p2p) which is almost
totally dead. While partly this is because new alt.* groups get
miserable propagation and partly because computer companies tend to
avoid alt completely and partly due to my lack of promotion of the
group, I think a major reason is the lack of interest in discussing
this subject on Usenet.
PS. If you do a second RFD I suggest you at least post a pointer in
alt.internet.p2p--if not add it to your distribution.
--
Jonathan Grobe
Jay Denebeim
Rajkumar Buyya <rajk...@csse.monash.edu.au> wrote:
>Given so many arguments, I also favour comp.distributed, but two
>other levels like:
>comp.distributed.grid
>comp.distributed.p2p
You don't have the traffic for that. If you go this route IMO they'll
both wither due to lack of 'critical mass'. Therefore if you go with
one group that has a reasonable name I'll abstain, but if you go with
two I'll vote against it.
Dave DiNucci
>
> In article <3BCADE7D...@csse.monash.edu.au>,
> Rajkumar Buyya <rajk...@csse.monash.edu.au> wrote:
>
> >Given so many arguments, I also favour comp.distributed, but two
> >other levels like:
>
> >comp.distributed.grid
> >comp.distributed.p2p
>
> You don't have the traffic for that. If you go this route IMO they'll
> both wither due to lack of 'critical mass'. Therefore if you go with
> one group that has a reasonable name I'll abstain, but if you go with
> two I'll vote against it.
Raj,
I, also, would suggest that it start with just one group and split later
if necessary. (Easier to work with the proof of too much traffic than
with the fear of not enough.) Splitting prematurely can lead not only
to problems with critical mass, but also unnecessary duplication and
over-categorization of discussions that would be more productive in a
broader setting. In fact, I imagine that some people would take issue
with the p2p/grid distinctions I've already made.
I can see how a name like "distributed" might get less backlash than my
preferred "collectives". (I have some expectation that the name
"distributed" will be interpreted by some as relating to distributed
COMPUTING only--i.e. less focus on other resource types, less focus on
issues that come out of the "collective" aspects rather than the
"distributed" ones--but a clear charter and the potential for future
splits should probably handle any confusion.)
Thanks for getting the ball rolling on this,
-Dave
-----------------------------------------------------------------
David C. DiNucci http://www.elepar.com Prog. Tools for Grid/P2P
Aahz Maruch
Rajkumar Buyya <rajk...@csse.monash.edu.au> wrote:
>
>It is nice to see many important issues and concerns which I was unable
>to raise them with other discussion groups. Many of Usnet leaders
>feel that 2nd level need to be concepts focus rather that implementation
>and also, I noticed that few other colleagues in discussion seem to support
>comp.distributed, it might be worth exploring it further. Something like:
> comp.distributed.grid
> comp.distributed.p2p
>
>We can have many sub level within p2p / grid depending on their relevance.
>All common things can be outside as they are need for both P2P and grid world.
Yup. However, start with just comp.distributed first, to see how much
traffic there is. If it starts up with tremendous volume, you'll be
able to split it in three months.
Rajkumar Buyya
Aahz Maruch wrote:
>
> In article <3BCADABD...@csse.monash.edu.au>,
> Rajkumar Buyya <rajk...@csse.monash.edu.au> wrote:
> >
> >It is nice to see many important issues and concerns which I was unable
> >to raise them with other discussion groups. Many of Usnet leaders
> >feel that 2nd level need to be concepts focus rather that implementation
> >and also, I noticed that few other colleagues in discussion seem to support
> >comp.distributed, it might be worth exploring it further. Something like:
> > comp.distributed.grid
> > comp.distributed.p2p
> >
> >We can have many sub level within p2p / grid depending on their relevance.
> >All common things can be outside as they are need for both P2P and grid world.
>
> Yup. However, start with just comp.distributed first, to see how much
> traffic there is. If it starts up with tremendous volume, you'll be
> able to split it in three months.
I agree. That is what we have been thinking given so much of overlap and comments
from the group.
cheers
Raj
> --
> --- Aahz <*> (Copyright 2001 by aa...@pobox.com)
>
> Hugs and backrubs -- I break Rule 6 http://www.rahul.net/aahz/
> Androgynous poly kinky vanilla queer het Pythonista
>
> We must not let the evil of a few trample the freedoms of the many.
--
Rajkumar Buyya
unmoderated group comp.distributed
This is a formal Request For Discussion (RFD) for the creation of
world-wide unmoderated Usenet newsgroup comp.distributed. This is
not a Call for Votes (CFV); you cannot vote at this time. Procedural
details are below.
Newsgroup line:
comp.distributed Distributed Resource Sharing and Exploitation.
CHANGES from previous RFD:
This is an updated version of the previously submitted RFD for
comp.p2p-grid. It addresses many comments and concerns raised during
discussion on the earlier RFD including the recommendation of a new
name, comp.distributed.
RATIONALE: comp.distributed
Networks in general, and the internet specifically, have been
evolving, from star topologies of thin clients or dumb terminals
connected to central servers, to a collection of highly connected
nodes, many having significant compute, storage, and peripherals,
along with human presence. Likewise, internet tools and protocols
have evolved from being primarily a mechanism to "push" (via email)
or "pull" (via web-browser) untyped data, into supporting more
interactive, semantic, and bi-directional relationships. These
changes have prompted different communities to (re-)explore the
potential of sharing and exploiting collections of heterogeneous,
geographically distributed resources such as computers, data, people,
and scientific instruments in a secure and consistent manner, usually
lacking any central control or authority. These efforts are often
described with terms like "peer-to-peer" ("p2p") and "grids", and
can serve to virtualize enterprizes by blurring the significance of
physical location.
Different communities tend to focus on different varieties of
resources, different overall objectives and constraints, and different
degrees of permanence of the resource collectives. For example,
"grid" communities will often consider large, semi-permanent (though
dynamically constituted) collections of world-class resources that can
be accessed much as utilities, to provide unprecedented capabilities
that enable, for example, large-scale problems in science, engineer-
ing, and commerce. "p2p" communities, on the other hand, often seek
on-demand temporary relationships between everyday personal computers,
devices, and peripherals "at the edge of the network", that help to
solve every-day problems of sharing, collaboration, and computing in
more efficient, convenient, and economical ways. Similar relation-
ships have been explored over time in areas related to human collabor-
ation, distributed data bases, distributed search, parallel and
distributed computing, web services, and hierarchical content delivery
networks.
In spite of these differences, all of these communities share a large
number of challenges as a direct result of attempting to effectively
and synergistically assemble and use these collectives of hetero-
geneous distributed resources. These challenges include:
* Lack of any central authority, leading to the potential unannounced
availability or withdrawal of resources, requiring fault tolerant
applications and complicating the discovery and scheduling of
resources.
* Heterogeneous resources, requiring methods to recognize and request
unique functionality when needed, while hiding unexploitable
resource differences behind consistent interfaces.
* Heterogeneous performance in those resources, prompting the use of
simulation and performance modeling to determine which resources to
use when.
* Heterogeneous requirements from both resource owners and end users
in terms of their objectives, quality of services, and computa-
tional economy.
* Unpredictable and dynamic network topology and properties,
requiring the ability to portably deal with differing latency and
bandwidth constraints (e.g. hiding latency while minimizing
overhead) and motivating quality of service (QoS) mechanisms.
* A complex and unpredictable concurrent environment, requiring
general approaches to program development that hide these features
while leveraging existing tools, languages, and techniques wherever
possible.
* A memory hierarchy that can extend to the memory and disk throughout
the collective, prompting a reconsideration of traditional data
storage and caching approaches.
* The potential presence of untrusted resources and/or actors,
requiring decentralized approaches to privacy, authorization,
authentication, anonymity, and the determination of levels of
acceptable risk associated with different operational modes.
* Achieving return on investment for both resource users and
providers, requiring approaches for auditable accounting and re-
imbursement as well as the consideration of cost/price as a resource
selection parameter.
* Impediments to connectivity, including firewalls and oversubscribed
scarce network resources (such as dial-in modems, and IP addresses
shared through network address translation/IP masquerading).
* Cross-organizational IT involvement, requiring flexible and
politically acceptable policies, procedures, and management tools.
* Evaluating and proposing mechanisms and policies for the protection
of intellectual property in an environment explicitly designed to
facilitate instant sharing.
* Understanding and exploiting the potential value of these resource
collectives, including effective collaboration strategies,
integration of mixed resource types into problem solving
environments, novel application areas and solution approaches
enabled by this environment, and the use of automated agents.
Already, international academic and commercial forums like:
* Global Grid Forum: <http://www.gridforum.org>
* Peer to Peer Computing WG: <http://www.p2pwg.org>
* Universal Plug-n-Play Forum <http://www.upnp.org>
* New Productivity Initiative <http://www.newproductivity.org>
have evolved to create standards and protocols for inter-operability
between heterogeneous systems providing virtual services. Recently,
infrastructure projects like the NSF Distributed TeraScale Facility
have focused even more attention, and include involvement from several
companies. Many computer and/or software vendors, large and small,
have recently announced specific projects or general priorities into
p2p and/or grids, including IBM, Intel, DSTC, Sun, and Microsoft.
Some details on these and other projects can be found at:
* http://www.gridcomputing.com/
* http://www.computer.org/dsonline/gc/index.htm
* http://www.peertal.com/
* http://www.nwfusion.com/
* http://www.peerintelligence.com/
* http://www.openp2p.com/
Although over 20 discussion mailing lists operated by individuals or
institutions exist, they are generally intended for discussion of
specific group priorities, and strongly segregate p2p and grid
communities, even when addressing similar issues. Another concern is
that mailing lists are likely to generate large volume of email for
members; therefore, moderators will often discourage use of these
lists for general or controversial discussion, and many prospective
participants feel discouraged from subscribing, do not become members,
and do not join important topical discussions. We believe that having
a newsgroup where people can participate in discussions of their own
choosing, when they want, without getting swamped with emails, will
help overcome these limitations and will encourage discussion and
dissemination without the need of explicit membership. While some
existing newsgroups, like comp.parallel and comp.sys.super, touch on
some specialized aspects of this topic, and will continue to do so,
this new group will serve as a focal point for considering the inter-
relationships, interactions, and synergies when combining these
separate technologies.
Strategy for publicising the comp.distributed newsgroup:
The formation of the comp.distributed newsgroup will be publicised
through the following channels (but not limited to):
* IEEE DS Online,
* Global Grid Forum,
* P2P WG,
* Grid Infoware,
* IEEE/ACM conferences:
* CCGRID'xy: <http://www.ccgrid.org/>,
* GRID'xy: <http://www.gridcomputing.org/>,
* Yahoo Group on gridcomputing as part of GridInfoware.
* IEEE Task Force on Cluster Computing (TFCC)
* Newsgroups such as comp.parallel
END RATIONALE.
CHARTER: comp.distributed
Although the name "comp.distributed" has been chosen due to its
familiarity and convenience, the group is to be broader than just
those topics traditionally regarded as "distributed computing".
Specifically, topics are to include any unique issues relating to
the creation and exploitation of collectives of geographically
distributed and potentially heterogeneous resources such as computers,
data/information sources, peripherals, instruments, and humans.
Appropriate areas of discussion in this context would include (but
are not limited to):
* discovering, scheduling/brokering, and accessing remote resources
* exploitation of heterogeneous resources
* resource management, scheduling, and computational economy
* portable/adaptable communication substrates
* quality of service approaches
* portable program development tools, languages, techniques
* data management tools and techniques
* exploitation of distributed memory hierarchy
* decentralized security
* practical accounting, reimbursement, and business & revenue models
* overcoming impediments to wide-area connectivity
* cross-organizational policy issues and ways to address them
* mechanisms and policies for intellectual property
* programming tools, environments, and languages
* applications, collaboration, and distributed agents
* simulation and performance modelling
* comparisons of grid and p2p, and issues unique to each
* events, surveys, news and general announcements
It is expected that additional 3rd-level subgroups addressing some of
these topics or others may be created as dictated by the volume and
cohesiveness of resulting message traffic.
END CHARTER.
PROCEDURE:
This is a request for discussion, not a call for votes. In this phase
of the process, any potential problems with the proposed newsgroups
should be raised and resolved. The discussion period will continue for
a minimum of 21 days (starting from when the first RFD for this proposal
is posted to news.announce.newgroups), after which a Call For Votes
(CFV) will be posted by a neutral vote taker. Please do not attempt to
vote until this happens.
All discussion of this proposal should be posted to news.groups.
This RFD attempts to comply fully with the Usenet newsgroup creation
guidelines outlined in "How to Create a New Usenet Newsgroup" and "How
to Format and Submit a New Group Proposal". Please refer to these
documents (available in news.announce.newgroups) if you have any
questions about the process.
END PROCEDURE.
DISTRIBUTION: comp.distributed
This RFD has been posted to the following newsgroups:
news.announce.newgroups
news.groups
comp.arch
comp.parallel
comp.parallel.pvm
comp.parallel.mpi
comp.sys.super
comp.client-server
and to the following mailing lists:
<gridco...@yahoogroups.com>
<decentra...@yahoogroups.com>
<PTPWG-DI...@PEACH.EASE.LSOFT.COM>
END DISTRIBUTION.
Proponent: Rajkumar Buyya <rajk...@csse.monash.edu.au>
Proponent: David C. DiNucci <da...@elepar.com>
ru.ig...@usask.ca
> unmoderated group comp.distributed
>Newsgroup line:
>comp.distributed Distributed Resource Sharing and Exploitation.
>RATIONALE: comp.distributed
>Networks in general, and the internet specifically, have been
[snip]
> enabled by this environment, and the use of automated agents.
As long as it was, the above section was informative and I think
necessary if there is a potential problem with different camps
of this kind of computing. After all the rationale is partly
intended to convince potential readers that this group will
serve their purposes or needs.
But...
>Already, international academic and commercial forums like:
[snip list]
>have evolved to create standards and protocols for inter-operability
>between heterogeneous systems providing virtual services. Recently,
>infrastructure projects like the NSF Distributed TeraScale Facility
>have focused even more attention, and include involvement from several
>companies. Many computer and/or software vendors, large and small,
>have recently announced specific projects or general priorities into
>p2p and/or grids, including IBM, Intel, DSTC, Sun, and Microsoft.
>Some details on these and other projects can be found at:
[snip list]
Ok, after this posting I'm definitely adding to my sig. Real world
popularity and initiatives are of little concern here, and really
have less impact on the success or failure of a proposal, generally
speaking, than many proponents think. If the above vendors don't
bother with newsgroups, their company's involvement is not going
to mean a thing on the number of readers. If it is intended to
impress readers, that's all it does, since really the only folks
that would be impressed by the above are the ones already interested
in the topic. It won't (shouldn't) move the folks that aren't
interested in the topic because they aren't supposed to vote.
I'd suggest throwing out the above paragraph.
>Although over 20 discussion mailing lists operated by individuals or
>institutions exist, they are generally intended for discussion of
If you are going to cite a list, this is the one that might
be of interest, along with readership stats of those who are
interested in the new group. If you can't demonstrate these
mailing lists actively support your initiative, there's no
point in mentioning them here, either.
>specific group priorities, and strongly segregate p2p and grid
>communities, even when addressing similar issues. Another concern is
>that mailing lists are likely to generate large volume of email for
>members;
Tell us how many is "large volume". No one should take that
term seriously here without quantification.
>therefore, moderators will often discourage use of these
>lists for general or controversial discussion, and many prospective
>participants feel discouraged from subscribing, do not become members,
>and do not join important topical discussions. We believe that having
>a newsgroup where people can participate in discussions of their own
>choosing, when they want, without getting swamped with emails, will
>help overcome these limitations and will encourage discussion and
>dissemination without the need of explicit membership.
Really, the excuse given above is always true for any mailing
list, it's just a question of degree. I wouldn't mind seeing
some supporting evidence that this is really a problem, e.g. a
word from many of the mailing list maintainers. I'd really
hate to see a potentially usable group be populated by
malcontent rejects of mailing lists creating what should really
be a .advocacy group, for example (though in that case, if
there are that many malcontents the topic should have enough
interested usenet readers, or be a total washout).
New paragraph.
>While some
^^^^^
delete
>existing newsgroups, like comp.parallel and comp.sys.super, touch on
>some specialized aspects of this topic, and will continue to do so,
>this new group will serve as a focal point for considering the inter-
^
New sentence: "However, this new group will allow more general
discussions, as well as serve as a focal..."
>relationships, interactions, and synergies when combining these
>separate technologies.
And the above paragraph should come BEFORE your mailing list
section.
>Strategy for publicising the comp.distributed newsgroup:
>The formation of the comp.distributed newsgroup will be publicised
>through the following channels (but not limited to):
[list snipped]
I don't see the point in this list. Once the group is created,
it out of the hands of us news.group folks, and the users of
the new group are free not to advertise it. Frankly, if it
passes, it should really be beyond the need for any hard selling.
I'd say snip this "publicizing" list.
>END RATIONALE.
>CHARTER: comp.distributed
>Although the name "comp.distributed" has been chosen due to its
>familiarity and convenience, the group is to be broader than just
^^^^^^^^^
the discussion in this
unmoderated newsgroup
>those topics traditionally regarded as "distributed computing".
[snip]
>END CHARTER.
[snip]
>DISTRIBUTION: comp.distributed
>This RFD has been posted to the following newsgroups:
[list snipped]
Have the proponents determined how many or which ISPs are blocking
their RFD (due to G5 spam filtering), and how many interested
folks (i.e. potential voters) might be affected?
>END DISTRIBUTION.
ru
--
My standard proposals rant:
Quality, usefulness, or merit of a topic is more or less irrelevant
in creating a new Big-8 newsgroup. Usenet popularity is the primary
consideration.
ba...@digital-marketplace.net
<snip>
>
> Have the proponents determined how many or which ISPs are blocking
> their RFD (due to G5 spam filtering), and how many interested
> folks (i.e. potential voters) might be affected?
<*sighes*> I can not speak for proponents but the last two posts up
thread are reported unavailable/exprired so at a guess I would say many
(perhaps some or all) having UNS feeds did not see the 2nd RFD.
>
> >END DISTRIBUTION.
>
> ru
>
> --
> My standard proposals rant:
> Quality, usefulness, or merit of a topic is more or less irrelevant
> in creating a new Big-8 newsgroup. Usenet popularity is the primary
> consideration.
--
news:alt.pagan FAQ at http://www.dmcom.net/bard/altpag.txt
news:alt.religion.wicca FAQ at http://www.dmcom.net/bard/arwfaq2.txt
Thomas Gagne
--
.tom
Shrisha Rao
Rajkumar Buyya <rajk...@csse.monash.edu.au> wrote:
> REQUEST FOR DISCUSSION (RFD)
> unmoderated group comp.distributed
>
>This is a formal Request For Discussion (RFD) for the creation of
>world-wide unmoderated Usenet newsgroup comp.distributed. This is
Gets my vote, for what it counts. I'm not too concerned about the
wording of the charter, etc., as those are chimerical issues and have
nothing to do with the actual functioning of an unmoderated newsgroup.
However, it remains an open question whether there are enough votes in
favor to pass the group.
Regards,
Shrisha Rao
absyst...@techemail.com
Ad...@cwcom.net
afr...@ohio.usa.com
alt...@omplace.net
a...@ddsl.net
an...@iibsi.com
anj...@technosoftusa.net
anno...@murasu.com
ar...@alloymail.com
at...@vsnl.com
bap...@yahoo.com.ar
bha...@aitcusa.com
bim...@worldnet.att.net
bookpub...@tm.net
bpit...@beachaccess.com.au
carala...@theinternetdepot.com
car...@kaazoo.com
Ca...@khojhyderabad.com
catz...@greynetwork.net
ch...@elec.superprism.net
chr...@ab7.smptserver.com
comm...@bargaindog.com
con...@ccavenue.com
content-m...@high-tech-communications.com
crys...@epost.de
c...@expokey.com
dal...@boardx.co.kr
dbst...@orgio.net
devan...@icenet.net
dma...@rtsi-nj.com
down...@attglobal.net
dsd...@hanmail.net
dsi...@covenant-isp.com
eb...@eudoramail.com
edi...@tm.net
educ...@edventures.com
Er...@casm.insa-lyon.fr
esto...@lopermanagement.com
expor...@indiatimes.com
feed...@khojhyderabad.com
free...@freechal.com
fxcbo...@sunpoint.net
gamb...@vsnl.com
glo...@statue.com
GOALS...@MAIL.COM
go...@biz.com
gl...@zxmail.com
har...@econosoftinc.com
har...@haruwen.com
heal...@healthna.net
heal...@healthna.net
hinduis...@about.com
hu...@free4all.com
IndiaE...@mediaone.net
ind...@pujaprasad.com
i...@strictly4men.nl
in...@domainstop.org
in...@helderlicht.com
in...@honorpub.com
in...@iskconinteractive.com
in...@jazzconnect.com
inf...@americassenior.com
i...@ntlworld.com
jack...@iol.it
jimt...@cyada.com
joebla...@arabia.com
jpd...@hanmail.net
kai...@bigfoot.de
ka...@myrealbox.com
lainfo...@publicist.com
la...@acconsulting.net
lbac...@chartermi.net
lea...@webrizon.com
lindamr...@home.com
li...@honorpub.com
Lu...@VisaPro.com
ma...@oldguardpainters.com
mail...@expokey.com
market...@admin.gmcc.ab.ca
merchants...@juno.com
mik...@qudsmail.com
mnj...@integrandsystems.com
msu-s...@its.tc.msu.edu
mycasin...@mail.com
nata...@ms43.url.com.tw
na...@uole.com
nb...@nbit.com
NCSA20...@ombramarketing.com
Net...@pnonline.com
ne...@abdirectory.allbusiness.com
ne...@bluecom.com
newsl...@nriworld.com
nito...@networksstatus.com
nobett...@inter.net-zet.com
nort...@eurosport.com
ntri...@chenoainc.com
pasto...@ieg.com.br
payf...@indiatimes.com
perfec...@indiatimes.com
plep...@arabia.com
praneet...@proudindian.net
pr...@cemetrix.com
priy...@btechincorp.com
prod...@keytochina.net
prod...@keytochina.com
prom...@yapalot.com
raj...@enviteccorp.com
r...@enviteccorp.com
res...@cybercomresources.com
rks...@hanmail.net
rkse...@hanmail.net
sad...@mail.com
sai...@biz.com
sa...@phimkiemhiep.com
sa...@enventia.com
sa...@hustore.com
sa...@lilihouse.com
sa...@montageinc.net
s...@unifysolutions.com
sara...@networksstatus.com
sat...@quantumsoftech.com
sb...@80-20initiative.net
sdan...@prontomail.com
sean...@greynetwork.net
sea...@vsnl.com
se...@ecomputertech.com
shi...@finitestates.com
silva...@silvatravel.com
si...@magiclynk.com
sni...@pacbell.net
sol...@solnet-ar.com
solo...@post.com
speci...@myrealbox.com
s...@eliteesi.com
srin...@aol.com
su...@eliteesi.com
sup...@pujaprasad.com
sy...@orgio.net
sz...@163.com
themaya...@compuserve.com
tpn...@indiatimes.com
van...@seekercenter.net
vaness...@freeze.com
ve...@marlabs.com
ve...@rol.it
vin...@alliedgroups.com
vjar...@rtsi-nj.com
vo...@vsnl.net
vor...@cardtown.com
w.ro...@worldnet.att.net
webm...@10topgifts.com
webm...@diehardindian.com
webm...@hindu.org
webm...@keytochina.net
webm...@mediadial.com
westch...@uole.com
winw...@dreamwiz.com
Workin...@eurosport.com
workin...@geek.com
your...@uol.com.ar
yourp...@rediffmail.com
zcor...@vpm.com
ru.ig...@usask.ca
>I'm in favor of such a newsgroup.
Why? This is a discussion phase, add your 2-bits worth to it.
It might make more sense out of the proposal.
ru
--
My (updated) standard proposals rant:
Quality, usefulness, merit, or non-newsgroups popularity of a topic
David C. DiNucci
> >RATIONALE: comp.distributed
>
> >Networks in general, and the internet specifically, have been
> [snip]
> > enabled by this environment, and the use of automated agents.
>
> As long as it was, the above section was informative and I think
Thanks. We also questioned the length, but felt the expense was worth
the potential gain.
> But...
>
> >Already, international academic and commercial forums like:
> [snip list]
> >have evolved to create standards and protocols for inter-operability
> >between heterogeneous systems providing virtual services. Recently,
> >infrastructure projects like the NSF Distributed TeraScale Facility
> >have focused even more attention, and include involvement from several
> >companies. Many computer and/or software vendors, large and small,
> >have recently announced specific projects or general priorities into
> >p2p and/or grids, including IBM, Intel, DSTC, Sun, and Microsoft.
> >Some details on these and other projects can be found at:
> [snip list]
>
> Ok, after this posting I'm definitely adding to my sig. Real world
> popularity and initiatives are of little concern here, and really
> have less impact on the success or failure of a proposal, generally
> speaking, than many proponents think. If the above vendors don't
> bother with newsgroups, their company's involvement is not going
> to mean a thing on the number of readers. If it is intended to
> impress readers, that's all it does, since really the only folks
> that would be impressed by the above are the ones already interested
> in the topic. It won't (shouldn't) move the folks that aren't
> interested in the topic because they aren't supposed to vote.
> I'd suggest throwing out the above paragraph.
I see your point, but maybe we just didn't follow through? That is,
having companies and government institutions investing significant money
into this field strongly suggests lots of interest, and lots of interest
is necessary to generate quality message traffic on a newsgroup.
Interest alone may not be sufficient to justify a newsgroup, but it is
certainly necessary. And if the number of people involved in the field
can be shown to grossly outnumber those involved with existing lists,
and if the "barriers to entry" for reading or posting to a newsgroup is
less than to a mailing list, maybe this would provide some evidence to
the voters of potential use of the newsgroup.
Also, some of the projects are commercial, but some aren't (e.g.
National Science Foundation, NASA, etc.). And I am not so quick to
dismiss commercial aspects either. I would be more than happy if this
group reflected the same sort of professionalism and signal/noise ratio
present on comp.arch, for example, and most of the posters there
work(ed) for the most dominant companies in their field, even if the
companies themselves aren't necessarily considered to be involved with
newsgroups.
> >Although over 20 discussion mailing lists operated by individuals or
> >institutions exist, they are generally intended for discussion of
>
> If you are going to cite a list, this is the one that might
> be of interest, along with readership stats of those who are
> interested in the new group. If you can't demonstrate these
> mailing lists actively support your initiative, there's no
> point in mentioning them here, either.
OK. (This was left in from the previous draft.)
> >specific group priorities, and strongly segregate p2p and grid
> >communities, even when addressing similar issues. Another concern is
> >that mailing lists are likely to generate large volume of email for
> >members;
>
> Tell us how many is "large volume". No one should take that
> term seriously here without quantification.
Both the volume generated and the definition of "large" differs
considerably from list to list. For example, this morning the
decentralization list at yahoo was generating up to eight messages an
hour, and nobody was complaining--perhaps most people go to the web to
read, rather than receiving the messages. On the other hand, an
individual Global Grid Forum list may have fewer than 3 posts per week,
and I've had two moderators ask me to "slow down" or take it "off line"
when averaging fewer than 4 posts per day. The expectations of the
target audience, as well as tools to sift through posts, needs to be
accounted for, I guess. (Newsgroups, of course, have a wide variety of
such tools.)
> >therefore, moderators will often discourage use of these
> >lists for general or controversial discussion, and many prospective
> >participants feel discouraged from subscribing, do not become members,
> >and do not join important topical discussions. We believe that having
> >a newsgroup where people can participate in discussions of their own
> >choosing, when they want, without getting swamped with emails, will
> >help overcome these limitations and will encourage discussion and
> >dissemination without the need of explicit membership.
>
> Really, the excuse given above is always true for any mailing
> list, it's just a question of degree. I wouldn't mind seeing
> some supporting evidence that this is really a problem, e.g. a
> word from many of the mailing list maintainers. I'd really
> hate to see a potentially usable group be populated by
> malcontent rejects of mailing lists creating what should really
> be a .advocacy group, for example (though in that case, if
> there are that many malcontents the topic should have enough
> interested usenet readers, or be a total washout).
I personally am not at all concerned about having too many "malcontent
rejects", and some might even consider me one (though I helped plan the
NASA Info Power Grid project, and am supportive of the goals of it and
the GF and P2P WG). Either the mailing lists are satisfying everyone's
needs, in which case there is no need for a newsgroup, or they are not.
Our goal here is to show the need for a newsgroup, not to denigrate the
validity or utility of the mailing lists. The lists are what they are.
If the list maintainers should have something to say on the subject
here, we'll be happy to take it into consideration and/or respond, but I
don't think we should see the absence of their involvement as either
positive or negative, and the vast majority of the voters and eventual
posters will not be list maintainers.
> New paragraph.
> ...
Thanks for the recommended clarifications.
> >Strategy for publicising the comp.distributed newsgroup:
>
> >The formation of the comp.distributed newsgroup will be publicised
> >through the following channels (but not limited to):
>
> [list snipped]
>
> I don't see the point in this list. ...[omitted]...
> I'd say snip this "publicizing" list.
I left it in from the earlier CFD, and thought it was recommended. I'd
be happy to snip it.
> >DISTRIBUTION: comp.distributed
>
> >This RFD has been posted to the following newsgroups:
>
> [list snipped]
>
> Have the proponents determined how many or which ISPs are blocking
> their RFD (due to G5 spam filtering), and how many interested
> folks (i.e. potential voters) might be affected?
Should we (proponents) be involved in this? I can see it in all of the
listed newsgroups where I sit (i.e. on Earthlink). Other than that, I
don't really understand the problem, and don't know what I could/should
do to either research it or remedy it.
Thanks for the feedback,
-- Dave
-----------------------------------------------------------------
David C. DiNucci http://www.elepar.com Prog. Tools for grid/p2p
ba...@digital-marketplace.net
>
> ru.ig...@usask.ca wrote:
<snip>
> >
> > Have the proponents determined how many or which ISPs are blocking
> > their RFD (due to G5 spam filtering), and how many interested
> > folks (i.e. potential voters) might be affected?
>
> Should we (proponents) be involved in this? I can see it in all of the
> listed newsgroups where I sit (i.e. on Earthlink). Other than that, I
> don't really understand the problem, and don't know what I could/should
> do to either research it or remedy it.
Okay you should be a little concerned about distribution, if G5 is
preventing half of the readers being aware of your proposal they will
not think apout it or comment on it. If thr CFV gets sent out to the
same list of groups perhaps have your potential voters will not know the
vote is taking place.
As for determining how well the distribution worked (I will remind you,
I did not recieve the RFD) would be posting a question to the groups
(either one at a time or about 4 groups at a time) and ask if the isers
saw the RFD posting, in effect a straw poll to find out how many NSPs
might have filtered out the RFD due to excessive cross posts.
As for fixing the problem if it appears that many did not see it, the
answer is to post a pointer to the RFD copy on a web site. This is
premitted for an RFD (last time I checked), however is not premitted for
a CFV. During voting you can still prove a pointer to the CGV and
often in the pointer a contact is listed to contact the votetaker if
they can not get the CFV from n.a.n, n.g or the other groups that this
RFD went to.
You want most (if not all) in the related groups to know about this
discussion, this is the best time to apply fixes and address concerns.
It also ibcreases your boter base which tends to increase the chance of
passing. You will need 107 eo 125 Yes votes if there is not raised
major concerns about the name space, charter etc.
I so far have nor seen any major concerns, as fae as I can tell language
has been cleared up and reorganizied. However if only 50 people who
want the group knows about this proposal (or will see the CFV) there is
no way the group will pass.
>
> Thanks for the feedback,
> -- Dave
> -----------------------------------------------------------------
> David C. DiNucci http://www.elepar.com Prog. Tools for grid/p2p
--
ru.ig...@usask.ca
>ru.ig...@usask.ca wrote:
[snip]
>> Ok, after this posting I'm definitely adding to my sig. Real world
>> popularity and initiatives are of little concern here, and really
>> have less impact on the success or failure of a proposal, generally
>> speaking, than many proponents think. If the above vendors don't
>> bother with newsgroups, their company's involvement is not going
>> to mean a thing on the number of readers. If it is intended to
>> impress readers, that's all it does, since really the only folks
>> that would be impressed by the above are the ones already interested
>> in the topic. It won't (shouldn't) move the folks that aren't
>> interested in the topic because they aren't supposed to vote.
>> I'd suggest throwing out the above paragraph.
>I see your point, but maybe we just didn't follow through? That is,
>having companies and government institutions investing significant money
>into this field strongly suggests lots of interest,
No, it doesn't, not the kind of interest needed. You need USENET
interest. The kind of interest you allude to typically has little
correlation on the degree of interest at the USENET level. It is
quite possible that it is even anti-correlated for some topics.
That's why when you make statements of this sort, you need to
qualify it with usenet related information.
>and lots of interest
>is necessary to generate quality message traffic on a newsgroup.
>Interest alone may not be sufficient to justify a newsgroup, but it is
>certainly necessary.
Actually, usenet interest alone is the only justification for a
Big-8 newsgroup.
>And if the number of people involved in the field
>can be shown to grossly outnumber those involved with existing lists,
>and if the "barriers to entry" for reading or posting to a newsgroup is
>less than to a mailing list, maybe this would provide some evidence to
>the voters of potential use of the newsgroup.
Maybe. But now you are using the so-called "Field of Dreams"
rationale, which usually results in a failed proposal when that
is the primary rationale. You should do your best to justify
your proposal without ever mentioning this rationale.
>Also, some of the projects are commercial, but some aren't (e.g.
>National Science Foundation, NASA, etc.). And I am not so quick to
>dismiss commercial aspects either.
I don't see why commercial aspects should be dismissed. It's just
that it has little meaning in the Rationale for a group. It's
more relevant in the Charter.
>> >specific group priorities, and strongly segregate p2p and grid
>> >communities, even when addressing similar issues. Another concern is
>> >that mailing lists are likely to generate large volume of email for
>> >members;
>>
>> Tell us how many is "large volume". No one should take that
>> term seriously here without quantification.
>Both the volume generated and the definition of "large" differs
>considerably from list to list.
Let me rephrase. What YOU as a reader considers large may not
be compatible with a what our experience indicates will work
for a group. I suspect that there isn't a problem, but any
proposal should be more quantitative than qualitative. Some
of us get suspicious when the term "large" is used without
some sort of backing information. It's often a meaningless
term meant to bluff.
>For example, this morning the
>decentralization list at yahoo was generating up to eight messages an
>hour, and nobody was complaining--perhaps most people go to the web to
>read, rather than receiving the messages. On the other hand, an
>individual Global Grid Forum list may have fewer than 3 posts per week,
>and I've had two moderators ask me to "slow down" or take it "off line"
>when averaging fewer than 4 posts per day.
I can't say I'm impressed. That kind of traffic is fine for a tech
group, I guess. I was expecting much more than that. Some of
those lists are barely on the radar. Ok, so maybe my suspicion
that the lists have enough traffic was wrong. Start worrying.
Ok, the real concern is how many subscribers there are on that
list, volume is secondary. Or more relevant, how many are
considering reading the proposed group. Remember, "Usenet
popularity is the primary consideration". If you can use
that as your primary rational, nothing else matters (not the
spread of the topic, not the volume, not the difficulties
in the mailling lists, none of that would matter).
[snip]
>> I'd really
>> hate to see a potentially usable group be populated by
>> malcontent rejects of mailing lists creating what should really
>> be a .advocacy group, for example (though in that case, if
>> there are that many malcontents the topic should have enough
>> interested usenet readers, or be a total washout).
>I personally am not at all concerned about having too many "malcontent
>rejects", and some might even consider me one (though I helped plan the
>NASA Info Power Grid project, and am supportive of the goals of it and
>the GF and P2P WG).
Let me rephrase: I'd hate it if the supporters of the proposal
turned out to be creating the group to spite the mailing lists
and be a bitch forum for rejects of the mailing lists. I don't
expect this, but that's SORT OF the kind of picture that's
being painted if you twist it a bit. On the otherhand, there
hasn't been much to prevent me from twisting it like that, either.
>> I don't see the point in this list. ...[omitted]...
>> I'd say snip this "publicizing" list.
>I left it in from the earlier CFD, and thought it was recommended. I'd
>be happy to snip it.
The pointers list is recommended. That's totally different,
relevant because it deals with current activities for this
proposal that should be reported. The post creation publicizing
list has nothing to do with the group creation activities.
>> >DISTRIBUTION: comp.distributed
>>
>> >This RFD has been posted to the following newsgroups:
>>
>> [list snipped]
>>
>> Have the proponents determined how many or which ISPs are blocking
>> their RFD (due to G5 spam filtering), and how many interested
>> folks (i.e. potential voters) might be affected?
>Should we (proponents) be involved in this? I can see it in all of the
>listed newsgroups where I sit (i.e. on Earthlink).
Yes, you should be. We have already had a few reports from news.group
regular who have NOT seen it at their service. That's bound to mean
that some number of usenet readers interested in your topic will
never see the RFD or worse the CFV. Inquire in the newsgroups in
your distribution list, one group at a time; ask if there was
anyone that didn't or doesn't see the RFD and what service they
use if they didn't. Then show us the list of these services.
Don't crosspost to those groups, because that guarantees you get
no reports of missing RFDs. :)
>Other than that, I
>don't really understand the problem, and don't know what I could/should
>do to either research it or remedy it.
The problem is that when spam got pretty bad, some ISPs turned on
a feature of the newsgroup server that rejects any message that
has been crossposted to 5 or more (or is it more than 5?) newsgroups
(the G5 flag of the software). That is, if the Newsgroups: header
on your RFD has 5 or more groups listed, there is 100% chance that
some of your intended readership is not seeing it. Those ISPs need
to switch to one of the better features, or use an additional piece
of software that does a MUCH better job of spam filtering. I've been
led to believe using G5 is a bonehead idea for spam filtering by an
ISP, but we are still fighting the problem.
ru
--
My (updated) standard proposals rant:
Quality, usefulness, merit, or non-newsgroups popularity of a topic
ba...@digital-marketplace.net
>
> David C. DiNucci <da...@elepar.com> wrote:
> >ru.ig...@usask.ca wrote:
<snip>
>
> >listed newsgroups where I sit (i.e. on Earthlink).
>
> Yes, you should be. We have already had a few reports from news.group
> regular who have NOT seen it at their service. That's bound to mean
> that some number of usenet readers interested in your topic will
> never see the RFD or worse the CFV. Inquire in the newsgroups in
> your distribution list, one group at a time; ask if there was
> anyone that didn't or doesn't see the RFD and what service they
> use if they didn't. Then show us the list of these services.
> Don't crosspost to those groups, because that guarantees you get
> no reports of missing RFDs. :)
To elaborate, I do not know if UNS a.k.a. WUN is using G5 filter (I seem
to recall seeing some cross posts containing greater posts, however some
of the name space certainly bogus. e.g. alt.get.yourself.a-life that
might not be counted, I do know that this NSP certainly has a moderation
problem. Either G5 or moderation configuration problem could be reason
that no UNS/WUN users is seeing the RFD.
Proponents, based on what I know the level of concern should at least
start with any regs of the groups that are gated to Usenet via UNS/WUN.
That is if you see the NSP label I have or that has *goes to look for
one* usenetserver.com in it the odds are that those Usenet users
(readers and lurkers) did not see the RFD.
Aside to groupies: Did talk to tech (again) about getting a better
outsouced server, the odds look grim of my request being answered. In
some ways I think a UDP would be the best thing, I can work around that
should one occur. However last time I looked there has been no formal
requests and the few others that have proposed UDP in a.b.n-s-c were
more concerned about spam as opposed to misconfiguration.
>
<snip>
>
> ru
>
> --
> My (updated) standard proposals rant:
> Quality, usefulness, merit, or non-newsgroups popularity of a topic
> is more or less irrelevant in creating a new Big-8 newsgroup.
> Usenet popularity is the primary consideration.
David C. DiNucci
>
> David C. DiNucci <da...@elepar.com> wrote:
> >ru.ig...@usask.ca wrote:
> [snip]
> >> >communities, even when addressing similar issues. Another concern is
> >> >that mailing lists are likely to generate large volume of email for
> >> >members;
> >>
> >> Tell us how many is "large volume". No one should take that
> >> term seriously here without quantification.
>
> >Both the volume generated and the definition of "large" differs
> >considerably from list to list.
>
> Let me rephrase. What YOU as a reader considers large may not
> be compatible with a what our experience indicates will work
> for a group. I suspect that there isn't a problem, but any
> proposal should be more quantitative than qualitative. Some
> of us get suspicious when the term "large" is used without
> some sort of backing information. It's often a meaningless
> term meant to bluff.
Understood, but again, "large" in the context above (i.e. too large for
a manageable mailing list) is in the eyes of the beholders, who, in this
case, are the mailing list members and maintainers, and that view
differs from list to list, and is often not formal and/or published. In
general, I understand your concerns over the open term "large",
however.
> >For example, this morning the
> >decentralization list at yahoo was generating up to eight messages an
> >hour, and nobody was complaining--perhaps most people go to the web to
> >read, rather than receiving the messages. On the other hand, an
> >individual Global Grid Forum list may have fewer than 3 posts per week,
> >and I've had two moderators ask me to "slow down" or take it "off line"
> >when averaging fewer than 4 posts per day.
>
> I can't say I'm impressed. That kind of traffic is fine for a tech
> group, I guess. I was expecting much more than that. Some of
> those lists are barely on the radar. Ok, so maybe my suspicion
> that the lists have enough traffic was wrong. Start worrying.
See above. Extrapolating these figures to interest in a newsgroup is a
different (and tricky) task, especially when postings to some of these
lists are constrained by various (sometime unpublished) guidelines by
the leads and/or members to maintain focus on group goals and minimize
distracting emails (just as we said in the RFD). We are certainly not
criticizing those guidelines--they may very well serve their purposes
perfectly. We just observe that they don't fit our needs in several
circumstances, and a newsgroup apparently would.
> Ok, the real concern is how many subscribers there are on that
> list, volume is secondary. Or more relevant, how many are
> considering reading the proposed group. Remember, "Usenet
> popularity is the primary consideration". If you can use
> that as your primary rational, nothing else matters (not the
> spread of the topic, not the volume, not the difficulties
> in the mailling lists, none of that would matter).
On the p2p side, decentralization@yahoogroups currently has 652 members.
It looks like probably a few dozen people have posted to
zgp.org/pipermail/p2p-hackers, maybe a dozen at alt.internet.p2p, and
somewhat over 30 to the "bluesky" newsgroup at
news://franklin.oit.unc.edu/bluesky. There may be significant overlap
in the posters in these groups/lists, though I'm also sure that many
people, like me, have never been aware of many of them, and I'm surely
missing many more. Somewhat related newsgroups like alt.napster,
alt.gnutella, and alt.sci.seti remain active, with large numbers of
posters. On the grid side, the GGF counted 340 people at a recent
meeting, and I believe that virtually all of these people (as well as
significantly many others around the world, like me, who did not make
the meeting) are on at least one of their mailing lists (with relatively
little overlap with the P2P lists, from my personal observations).
I admit to being a bit confused by what constitutes acceptable evidence
of Usenet popularity. You now appear willing to consider subscribers to
these mailing lists as potential Usenet users, which is good. I am
hoping to avoid a "catch 22" situation before a CFV, where we are
expected to show that there are sufficient "Usenet people" who would
read or post specifically to comp.distributed, where the definition of a
"Usenet person" is one who already reads or posts to Usenet. (That is,
we wouldn't even be suggesting a new newsgroup if there were already
sufficient places on Usenet for these "Usenet people" to post about
comp.distributed topics--and measuring non-existent posts is difficult!)
> [snip]
> >> I'd really
> >> hate to see a potentially usable group be populated by
> >> malcontent rejects of mailing lists creating what should really
> >> be a .advocacy group, for example (though in that case, if
> >> there are that many malcontents the topic should have enough
> >> interested usenet readers, or be a total washout).
>
> >I personally am not at all concerned about having too many "malcontent
> >rejects", and some might even consider me one (though I helped plan the
> >NASA Info Power Grid project, and am supportive of the goals of it and
> >the GF and P2P WG).
>
> Let me rephrase: I'd hate it if the supporters of the proposal
> turned out to be creating the group to spite the mailing lists
> and be a bitch forum for rejects of the mailing lists. I don't
> expect this, but that's SORT OF the kind of picture that's
> being painted if you twist it a bit. On the otherhand, there
> hasn't been much to prevent me from twisting it like that, either.
If you'd hate it and don't expect it, then it's almost certainly not
worth discussing. Since it is an unmoderated group, we obviously can't
control what discussion will take place anyway, but "bitching" is not
listed in the charter, and I haven't seen (or said) anything to indicate
that complaining will occur. Nor do I understand what complainers would
hope to gain, especially since people posting to the newsgroup will
often likely be the same people (or work with people) posting to the
lists.
We have suggested that existing lists do not meet all the needs in this
field. This is the motivation for creating a newsgroup, not for
complaining. The fact that we (the proponents) have recognized the
potential value of a big-8 newsgroup on this topic from nearly opposite
ends of the earth suggests to me that our perceptions are not isolated.
I think this could be seen as a criticism of existing lists only for
people who insist that existing lists are intended to all the needs
listed in our charter and rationale--and I don't know those people.
Thanks again,
-- Dave
-----------------------------------------------------------------
David C. DiNucci Elepar Tools for portable grid,
da...@elepar.com http://www.elepar.com parallel, distributed, &
503-439-9431 Beaverton, OR 97006 peer-to-peer computing
David C. DiNucci
> > Ok, the real concern is how many subscribers there are on that
>
> On the p2p side, decentralization@yahoogroups currently has 652 members.
> It looks like probably a few dozen people have posted to
> zgp.org/pipermail/p2p-hackers, maybe a dozen at alt.internet.p2p, and
> somewhat over 30 to the "bluesky" newsgroup at
> news://franklin.oit.unc.edu/bluesky. There may be significant overlap
> in the posters in these groups/lists, though I'm also sure that many
> people, like me, have never been aware of many of them, and I'm surely
> missing many more. Somewhat related newsgroups like alt.napster,
> alt.gnutella, and alt.sci.seti remain active, with large numbers of
> posters. On the grid side, the GGF counted 340 people at a recent
> meeting, and I believe that virtually all of these people (as well as
> significantly many others around the world, like me, who did not make
> the meeting) are on at least one of their mailing lists (with relatively
> little overlap with the P2P lists, from my personal observations).
I omitted an important mailing list in the above,
PTPWG-DI...@PEACH.EASE.LSOFT.COM, which currently lists 552
subscribers. I apologize to that list for the omission, but it
apparently does not allow viewing of the archives without a password,
and I'm good at losing passwords (and somewhat wary of cookies). In
that context, it should be noted that the decentralization list and the
GGF lists do not require passwords/cookies to view their web archives,
so they may very well have many more readers than those listed as
members.
David C. DiNucci
I followed-up individually to each group (other than news.groups and
news.announce.newgroups), saying that there was an RFD, and providing a
pointer to a google web page where the original can be found. (I opted
not to get involved with trying to track down the G5 problem with tons
of return email, since I am probably not the best person for debugging
Usenet anyway, and I am satisfied as long as everyone sees either the
original post or my follow-up.)
I am still concerned about news.announce.newgroups, however. Would an
"unofficial" post like my follow-up likely make it through the
moderator? Recommendations welcome, else I'll use trial and error.
Thanks,
-- Dave
-----------------------------------------------------------------
ba...@digital-marketplace.net
<snip>
>
> I followed-up individually to each group (other than news.groups and
> news.announce.newgroups), saying that there was an RFD, and providing a
> pointer to a google web page where the original can be found. (I opted
> not to get involved with trying to track down the G5 problem with tons
> of return email, since I am probably not the best person for debugging
> Usenet anyway, and I am satisfied as long as everyone sees either the
> original post or my follow-up.)
<nods> At least all involved groups are aware of it, as to if they
click on the link is up to them. A repost of the RFD is permitted as
well, but there is no way of knowing if any would read it. As to G5
this is a concern and getting feedback would be nice to have. It might
be hard for a stranger to ask for it, but you certainly are not require
to do this. If your mail box is busy, I can well understand your
concern.
>
> I am still concerned about news.announce.newgroups, however. Would an
> "unofficial" post like my follow-up likely make it through the
> moderator? Recommendations welcome, else I'll use trial and error.
Well all properly gated posts will hit the moderator bot and yours most
likely would be rejected as a repost. However if you get an NSP
wothout moderation flag set you certainly can post to n.a.n just that
(hopefully) few will see it and it would not be considered valid
posting. Do not test n.a.n it is not the problem and might upset a few
people. You do not want to generate No votes from anywhere, you do need
your Yes votes.
>
> Thanks,
> -- Dave
> -----------------------------------------------------------------
> David C. DiNucci Elepar Tools for portable grid,
> da...@elepar.com http://www.elepar.com parallel, distributed, &
> 503-439-9431 Beaverton, OR 97006 peer-to-peer computing
--
news:alt.pagan FAQ at http://www.dmcom.net/bard/altpag.txt
news:alt.religion.wicca FAQ at http://www.dmcom.net/bard/arwfaq2.txt
news:news.groups (draft)FAQ at http://www.dmcom.net/bard/ngfaq.txt
David C. DiNucci
>
> In news.groups Rajkumar Buyya <rajk...@csse.monash.edu.au> wrote:
> > REQUEST FOR DISCUSSION (RFD)
> > unmoderated group comp.distributed
>
> > ...
> >therefore, moderators will often discourage use of these
> >lists for general or controversial discussion, and many prospective
> >participants feel discouraged from subscribing, do not become members,
> >and do not join important topical discussions. We believe that having
> >a newsgroup where people can participate in discussions of their own
> >choosing, when they want, without getting swamped with emails, will
> >help overcome these limitations and will encourage discussion and
> >dissemination without the need of explicit membership.
>
> Really, the excuse given above is always true for any mailing
> list, it's just a question of degree. I wouldn't mind seeing
> some supporting evidence that this is really a problem, e.g. a
> word from many of the mailing list maintainers. I'd really
> hate to see a potentially usable group be populated by
> malcontent rejects of mailing lists creating what should really
> be a .advocacy group, for example (though in that case, if
> there are that many malcontents the topic should have enough
> interested usenet readers, or be a total washout).
The RFD language above (and one of my earlier responses) seems to have
led to some unexpected conclusions, and I'd hate to see
misunderstandings lead to "no" votes, so I'm giving this another stab.
It is difficult to cite the merits of a newsgroup in the presence of so
many mailing lists without appearing to criticize mailing lists, but we
have no choice but to try. Our language above reflects, in part, our
personal experiences and observations when we and others have tested the
natural limits of mailing lists, but I agree that that is not so
important in itself. The bottom line is that, as good as these lists
are at achieving their individual goals, and as good as the list
maintainers are at working within the constraints that lists impose
(sometimes by trimming those goals), I don't think that any existing
lists attempt to meet all of the goals we mention in the
comp.distributed newsgroup charter and rationale. And, I don't think it
would be possible to create a mailing list to meet them. That's all that
really matters here.
The flip side is that a newsgroup could theoretically be viewed by some
as an unwelcome "competitor" for lists, but with more consideration,
this
seems unlikely. For example, I don't expect this newsgroup to have much
effect at all on the kind of traffic carried on lists (like the GGF
lists) with a specific "product" goal or administrative function, except
maybe to augment them by providing another source for ideas and/or a
venue for brainstorming. On the other hand, some of the more free-form
mailing lists and non-big-8 newsgroups on similar subjects may have been
happy with a group like comp.distributed in the first place, but went
with a list or alt group to avoid the overhead and time investment
related to big-8 newsgroup creation. (This is illustrated by Jonathan
Grobe, who created one such alt group and now appears supportive of
comp.distributed.) Of course, we wouldn't be proposing the newsgroup
unless we expected some people to prefer it to mailing lists, but if
other people really see that as a problem in itself--i.e. they hope to
maintain mailing list traffic by simply excluding all other options that
members might prefer--we can only hope that they will be outvoted by
more freedom-loving people. :-)
So, yes, I'd also be happy to see a word from mailing list maintainers
and members, or anybody else. If anyone has issues, we want to do our
best to address them now rather than just see them appear as unexplained
"no"
votes after a CFV.
-- Dave
--
David C. DiNucci
> > "unofficial" post like my follow-up likely make it through the
> > moderator? Recommendations welcome, else I'll use trial and error.
>
> Well all properly gated posts will hit the moderator bot and yours most
> likely would be rejected as a repost. However if you get an NSP
> wothout moderation flag set you certainly can post to n.a.n just that
> (hopefully) few will see it and it would not be considered valid
> posting. Do not test n.a.n it is not the problem and might upset a few
> people. You do not want to generate No votes from anywhere, you do need
> your Yes votes.
So, the way I read this, even though we have followed all of the rules
(as far as I can tell), there is no way for us to ensure that everyone
who keeps an eye on news.announce.newgroups will see that we posted an
RFD for this newsgroup (especially if we continue to follow the rules).
It looks to me, at the very least, like someone needs to update the
guidelines about newsgroup creation, but I don't even know what the new
guidelines would be. That is, if someone wants to post an RFD or CFV to
more than 5 groups, do you just tell them "no"? If proponents split it
the RFD into multiple "batches" to news.announce.newgroups, presumably
the duplicate posts will be flushed by the "mod bot" you mention. Maybe
the n.a.n moderator needs to be manually splitting posts into multiple
4-group batches?
I guess we'll need to clear this up before our CFV.
ba...@dmcom.net
>
<snips>
>
> So, the way I read this, even though we have followed all of the rules
> (as far as I can tell), there is no way for us to ensure that everyone
> who keeps an eye on news.announce.newgroups will see that we posted an
> RFD for this newsgroup (especially if we continue to follow the rules).
Yes that is correct if any interested person is using either G5 filter
or moderation setting in error NSP, they will not see the RFD or CFV
even though you followed the rules. It should be noted the rules were
relaxed some for your and the other recent RFD. This was done because
groupies (and perhaps others) have convinced some (many?) NSPs to either
give n.a.n an expemtion from a G% filter or to remove the G5 filter
totally (using different spam filters). So far I am the only reg of
this group that has not seen your RFD IIRC.
>
> It looks to me, at the very least, like someone needs to update the
> guidelines about newsgroup creation, but I don't even know what the new
> guidelines would be. That is, if someone wants to post an RFD or CFV to
> more than 5 groups, do you just tell them "no"? If proponents split it
> the RFD into multiple "batches" to news.announce.newgroups, presumably
> the duplicate posts will be flushed by the "mod bot" you mention. Maybe
> the n.a.n moderator needs to be manually splitting posts into multiple
> 4-group batches?
Well I am sure the moderators are concerned about this and will try to
find the best posible answer. I am not harmed much but nt seeing your
RFD because I see the daily reports and can look at the RFD by using
google. This is the advantage I have because I Know all the work
arrounds, including getting a ballot if I do not see the CFV. Your
voters however may not know enough to get arround the problem which is
the concern.
>
> I guess we'll need to clear this up before our CFV.
That was one reason ru asked if proponents had conducted a survey about
if the RFD was seen. I could try to do so as well, it depends a lot on
how busy the groups are and if any of them are moderated. The moderated
groups (if any) I most likely will need to know the e-mail submission
address. Oh I would not care to enter a hostile group either, however I
could delete sig to try to avoid that. Right now there is no way to
know how many Yes votes have not seen your RFD. Again I am sure any
UNS/WUN users did not see the post, pending verifcation this might
change. It might be only a few of my NSP's boxes or messed up, however
I do strongly suspect the filter and gateway is system wide.
Jay Denebeim
David C. DiNucci <da...@elepar.com> wrote:
>That is, if someone wants to post an RFD or CFV to more than 5
>groups, do you just tell them "no"?
No, they told them about the problem and left it up to the proponant.
There were always some who went ahead anyway.
Remember though G5 is *BAD* it's a lousy attempt at limiting spam.
There's much better ways to do it. G5 throws out way too much baby
with the bathwater and it doesn't work all that well either. In fact
it used to be G8, then they backed it down because it wasn't catching
very much spam.
Anyway, as Bard said, many people here have been trying to beat some
sense into the admins of sites. Aparently they've had some luck.
Jay
--
I'm looking for a job, for my resume please see:
http://www.deepthot.org:2001/denebeim.html
David C. DiNucci
>
> David C. DiNucci wrote:
> >
> <snips>
>
> > So, the way I read this, even though we have followed all of the rules
> > (as far as I can tell), there is no way for us to ensure that everyone
> > who keeps an eye on news.announce.newgroups will see that we posted an
> > RFD for this newsgroup (especially if we continue to follow the rules).
>
> Yes that is correct if any interested person is using either G5 filter
> or moderation setting in error NSP, they will not see the RFD or CFV
> even though you followed the rules. It should be noted the rules were
> relaxed some for your and the other recent RFD. This was done because
> groupies (and perhaps others) have convinced some (many?) NSPs to either
> give n.a.n an expemtion from a G% filter or to remove the G5 filter
> totally (using different spam filters).
Perhaps I'm still thinking of old-fashioned newsgroup propagation
schemes, but will this only affect news servers having the G5 filter, or
could it also affect others "downstream" from those servers?
If I had realized that n.a.n was treated specially by this filter, I
might have been more prone to trying to help debug it. (That is, I
assumed that anyone could run an experiment by cross-posting to most any
5 groups, but that's evidently not so.) Still, I didn't expect
"debugging Usenet" to be added to my list of priorities (which is
already overflowing), so I do appreciate any work that others are doing
(or have done) to alleviate the problem.
> So far I am the only reg of
> this group that has not seen your RFD IIRC.
I'll take that as a good sign!
Thanks,
-Dave
ba...@dmcom.net
<snips>
>
> > this group that has not seen your RFD IIRC.
>
> I'll take that as a good sign!
Well in case you do not know, UNS/WUN provides outsoucing service for
ISPs like bell south (severval thousand users) so those might result in
a few lost votes. There appears to be another server that also has
n.a.n misconfigured so there might be some interested readers there.
Debuging the Internet is not a task I think any could do, unless well
paid, provided clubs, hammers and a very good protection force in place
(OH and of course the ability to configure servers in accordance with
the RFCs most current).
There was another RFD posted to n.a.n that I saw that was not gated
properly. I will try to find out more about the size of that ISP,
perhaps if I can not get current outsouced NSP working maybe I can fix
the other one.
I tried to find out how many users this NSP serves and was unable to do
so, UNS is one of the largest out soucing NSP, so the numbers could
perhaps reach 100,000. How many of the users are interested is not
posible for me to say, the rule of thumb has been about 10 % of
connented people know what Usenet is and 10% of those that read usenet
will post. There is no current way to know (track) who reads what
groups or if these percentages have changed. Using the straight 1%
there might be 1000 votes out there for you as any one can vote.
To groupies: I do see today that UNS web page does claim about 80
peers. Looking a little more their undated page this date on retention
they list misc.jobs.offered 68 days and I know that group was removed.
<sighes>
Kathy I. Morgan
> So, the way I read this, even though we have followed all of the rules
> (as far as I can tell), there is no way for us to ensure that everyone
> who keeps an eye on news.announce.newgroups will see that we posted an
> RFD for this newsgroup (especially if we continue to follow the rules).
What you could do is point them to the copy which will be appearing
shortly here in news.groups. I've posted a followup to your 2nd RFD,
quoting it in it's entirety so there will be a copy readily accessible.
I don't know what the Message ID will be, but it should be appearing at
the same time this message does.
--
Kathy
visist news:news.groups.reviews to read reviews of other newsgroups
help for new users of newsgroups at <http://www.aptalaska.net/~kmorgan/>
Good Net Keeping Seal of Approval at <http://www.gnksa.org/>
Kathy I. Morgan
for the benefit of those who failed to receive the original due to G5 or
other problems. Kathy
David C. DiNucci
>
> David C. DiNucci <da...@elepar.com> wrote:
>
> > So, the way I read this, even though we have followed all of the rules
> > (as far as I can tell), there is no way for us to ensure that everyone
> > who keeps an eye on news.announce.newgroups will see that we posted an
> > RFD for this newsgroup (especially if we continue to follow the rules).
>
> What you could do is point them to the copy which will be appearing
> shortly here in news.groups. I've posted a followup to your 2nd RFD,
> quoting it in it's entirety so there will be a copy readily accessible.
> I don't know what the Message ID will be, but it should be appearing at
> the same time this message does.
If I understand correctly, I can't do that--that is, I understand that
nothing will get through the news.announce.newgroups moderator except
for a new RFD or CFV, etc., so I can't post a pointer there to
anything. For all the other groups except news.announce.newgroups (and
this one), yes, I have already posted a pointer to the groups.google
version of the RFD (and I suppose that I could have just quoted the
original RFD in a follow-up as you did, but I didn't want to
accidentally trip some little-known clause in the rules, so I followed
the advice I had been given here).
David C. DiNucci
>
> I have no comment to make; I'm just quoting the message in its entirety
> for the benefit of those who failed to receive the original due to G5 or
> other problems. Kathy
>
> Rajkumar Buyya <rajk...@csse.monash.edu.au> wrote:
>
> > REQUEST FOR DISCUSSION (RFD)
> > unmoderated group comp.distributed
> >
> > This is a formal Request For Discussion (RFD) for the creation of
> > world-wide unmoderated Usenet newsgroup comp.distributed. This is
> > not a Call for Votes (CFV); you cannot vote at this time. Procedural
> > details are below.
Thanks for the repost--for those who didn't see the RFD before, and so
have been wondering what we've been talking about.
Although we would very much like to see more discussion, at least we
have seen very little negative discussion (in spite of our widespread
dissemination and followup pointers of the RFD), so we are preparing to
move on to the next step--i.e. a vote--hopefully within the next two
weeks. But, I want to make sure that it can go smoothly, so...
(a) I see that a CFV cannot occur within 10 days of the last RFD, so if
we need another RFD, we should make it soon. So far, the only
suggestions I have seen here have been for very minor wording changes,
re-ordering of words, omitting words, etc., and I don't know if these
justify another RFD, but if they do, I'd like to know now if possible.
(b) Of course, the G5 problem is waiting to bite us again on the CFV
(and the followup RFD, if one is required), and I don't know how to stop
it. I understand that the CFV is usually posted to the same groups as
the RFD, and in fact, we would like it to be posted to the same groups.
Since we see the problems up ahead, does anybody have suggestions on how
to steer around them?
Thanks,
-Dave
ru.ig...@usask.ca
>(b) Of course, the G5 problem is waiting to bite us again on the CFV
>(and the followup RFD, if one is required), and I don't know how to stop
>it. I understand that the CFV is usually posted to the same groups as
>the RFD, and in fact, we would like it to be posted to the same groups.
>Since we see the problems up ahead, does anybody have suggestions on how
>to steer around them?
I think what has been recommended up to now is to trim down the
distribution, and post NEUTRAL pointers to "the very low traffic"
news.announce.newgroups in the remaining groups. Not sure about
the timing, or if the n.a.n moderator handles that. Consult the
mentors on the wording of the pointer as well as the procedure.
This might not be an issue if you get no one in those groups telling
you they didn't see the RFD (in response to your followups).
ru
--
My (updated) standard proposals rant:
Quality, usefulness, merit, or non-newsgroups popularity of a topic
is more or less irrelevant in creating a new Big-8 newsgroup.
Jay Denebeim
David C. DiNucci <da...@elepar.com> wrote:
>(a) I see that a CFV cannot occur within 10 days of the last RFD, so if
>we need another RFD, we should make it soon. So far, the only
>suggestions I have seen here have been for very minor wording changes,
>re-ordering of words, omitting words, etc., and I don't know if these
>justify another RFD, but if they do, I'd like to know now if possible.
No, they don't. You can make the changes on the final charter you
send to be put on the CFV. That's the one that ends up being recorded
everywhere.
>(b) Of course, the G5 problem is waiting to bite us again on the CFV
>(and the followup RFD, if one is required), and I don't know how to stop
>it. I understand that the CFV is usually posted to the same groups as
>the RFD, and in fact, we would like it to be posted to the same groups.
>Since we see the problems up ahead, does anybody have suggestions on how
>to steer around them?
Yeah, you might be able to convince the UVV person into multi-posting
the CFV.
ru.ig...@usask.ca
>I admit to being a bit confused by what constitutes acceptable evidence
>of Usenet popularity.
Basically, either a) a lot of folks already posting to usenet on
the topic, b) a lot of folks who say they are reading usenet
postings on the topics, c) a lot of folks not on usenet who say
they want to read a newsgroup on the topic.
>You now appear willing to consider subscribers to
>these mailing lists as potential Usenet users, which is good.
I always have, as potential. But by no means do I believe they
will. As I've said before, we have had plenty of proposals
passed and failed in which the newsgroup turned out NOT to be
used, and the web board or mailing list being prefered. That
is why the mere presence of even readership on non-usenet
forums is pretty much meaningless without some kind of usenet
oriented qualification. That usually means surveying the mailing
list readership in this situation.
>I am
>hoping to avoid a "catch 22" situation before a CFV, where we are
>expected to show that there are sufficient "Usenet people" who would
>read or post specifically to comp.distributed, where the definition of a
>"Usenet person" is one who already reads or posts to Usenet.
It more of a question of determining the primary rationale.
If it is a question of reorganizing existing discussions, and
even increasing the scope, then the existence of mailing lists
is moot (too unreliable). If, on the otherhand, you don't
have enough usenet usage, you are going to want to demonstrate
that this proposal is being presented seriously by showing
that the readership really is out there ready to move into
newsgroups (that ceased to be a given about 2 to 3 years ago).
There are some folks that would consider the notion of mailing
list readers moving to newsgroups something of a joke, and
while they may be few, other small issues with a proposal can
collect enough similar minded folks to make your attempt
more difficult.
Finally, going out there and surveying the non-usenet population
will raise the awareness of those readers BEFORE the vote.
Really, this should have been done before the RFD, so that
more of them could participate in the current discussions.
There's even no guarantee that most of them have used newsgroups
before, and this might be the only time they find out how
to use newsgroups so that they can even vote. Regardless,
the CFV is DEFINITELY NOT the time you want to be trying
to get folks' attentions; there are restrictions on what
you (being a proponent) can actually post at that time.
>(That is,
>we wouldn't even be suggesting a new newsgroup if there were already
>sufficient places on Usenet for these "Usenet people" to post about
>comp.distributed topics--and measuring non-existent posts is difficult!)
Not true. There may well be places where comp.distributed topics
are discussed and on topic, but then it is your job to rationalize
why they are not adequate. I think you've done a fair bit of
that, as far as the topic space is concerned.
>ru.ig...@usask.ca wrote:
>> Let me rephrase: I'd hate it if the supporters of the proposal
>> turned out to be creating the group to spite the mailing lists
>> and be a bitch forum for rejects of the mailing lists. I don't
>> expect this, but that's SORT OF the kind of picture that's
>> being painted if you twist it a bit. On the otherhand, there
>> hasn't been much to prevent me from twisting it like that, either.
>If you'd hate it and don't expect it, then it's almost certainly not
>worth discussing.
But it is worth discussion because no one cares if I hate it,
and the unexpected has a nasty way of manifesting, especially
since I have no way of determining if my expectations are
real.
>Since it is an unmoderated group, we obviously can't
>control what discussion will take place anyway, but "bitching" is not
>listed in the charter, and I haven't seen (or said) anything to indicate
>that complaining will occur.
You sort of implied the possibility when you say folks opt
out of the mailing lists for various reasons. They have
reasons for complaining, some of which is simply due to
traffic, others for the ML administration, etc. This implies
some degree of disfranchisement, which at higher degrees is
not desirable. We've had our share of those here.
>Nor do I understand what complainers would
>hope to gain, especially since people posting to the newsgroup will
>often likely be the same people (or work with people) posting to the
>lists.
Hypothetically, the complainers can drive the MLers back to their
ML, leaving a flame/dead group (for which the regulars here bemoan the
lack of a way of cleaning up - that story's complicated).
Overall, what I'm saying is plug as many of the holes in your
proposal you can to reduce the number of NO votes you might
get from them. Did I mention anywhere that there are a fair
number of reasons (not all reasonable) that some folks vote
NO for? There's an FAQ on that, probably available at the
same place you got the other creation FAQs ("Why People Vote
NO", or something like that). Just remember that your
readers will be voting YES, and that folks that think your
group will damage their forum and news.group folks that
think your proposal is flawed (poor charter, poor name,
deceptive rationale) will vote NO. Thus, one of your jobs
is actually to reduce the NO votes.
ru
--
My (updated) standard proposals rant:
Quality, usefulness, merit, or non-newsgroups popularity of a topic
is more or less irrelevant in creating a new Big-8 newsgroup.
David C. DiNucci
So, in brief, you are willing to broaden the definition of "Usenet
people" to those who can be shown, through a survey, to be willing to
read or post to this newsgroup. I think that's exactly what a CFV is.
So, you are recommending a straw vote?
> >(That is,
> >we wouldn't even be suggesting a new newsgroup if there were already
> >sufficient places on Usenet for these "Usenet people" to post about
> >comp.distributed topics--and measuring non-existent posts is difficult!)
>
> Not true. There may well be places where comp.distributed topics
> are discussed and on topic, but then it is your job to rationalize
> why they are not adequate. I think you've done a fair bit of
> that, as far as the topic space is concerned.
My position has been both that (a) there are not sufficient on-topic
places on Usenet to accomodate most of the comp.distributed discussion,
and (b) even in those pockets where bits and pieces of the discussion
are on topic, they are so widely scattered as to not influence one
another (as they must to be productive).
> >Since it is an unmoderated group, we obviously can't
> >control what discussion will take place anyway, but "bitching" is not
> >listed in the charter, and I haven't seen (or said) anything to indicate
> >that complaining will occur.
>
> You sort of implied the possibility when you say folks opt
> out of the mailing lists for various reasons. They have
> reasons for complaining, some of which is simply due to
> traffic, others for the ML administration, etc. This implies
> some degree of disfranchisement, which at higher degrees is
> not desirable. We've had our share of those here.
You continue to read many things into our proposal that (a) we are not
saying, (b) I can't understand, and (c) I have already clarified to be
opposite to your implications (e.g. in my post yesterday morning). I
can't stop you from seeing what you like, but I think this portion of my
response is tailored more for you than for most other potential readers.
I believe that few "folks opt out of mailing lists". We said that there
are some sorts of *discussions* which don't work well on lists,
especially when those lists have been created with more specific goals.
In fact, I said that some people are likely to post to both mailing
lists and the newsgroup, and that the newsgroup can actually benefit the
lists in many cases. (I will now add, "and vice versa".)
I don't understand your contention that "they have reasons for
complaining". Are you suggesting that, if people normally use a
congested highway or eat at a restaurant with a strict dress code, and
then find a less congested alternate route or a more casual restaurant,
that that gives them reasons for complaining? And, moreover, that it is
justification for closing down the alternate route or casual
restaurant? On the other hand, given free reign, some discussion will
tend to evaluate and discuss the pros and cons of the surrounding
environment, and that can be a productive exercise.
You are also the only one claiming "some degree of disenfranchisement".
Are you suggesting that people will drop out of the Grid Forum or
Peer-to-Peer WG because there is a newsgroup in addition to their
mailing lists? Or are you suggesting that people will move discussion
and administration of the specific group projects off of those mailing
lists and into a general newsgroup? I can't imagine it. (Even in the
bizarre chance that something like this happened, if it helped the
overall functioning of the Forum or WG, they would likely see it as a
benefit!) It is also hard to think of someone feeling disenfranchised
if a mailing list that was created (let's say) for a particular standard
relating to p2p software issues doesn't want to accomodate more
wide-ranging topics that might (for now) relate more to grids.
> >Nor do I understand what complainers would
> >hope to gain, especially since people posting to the newsgroup will
> >often likely be the same people (or work with people) posting to the
> >lists.
>
> Hypothetically, the complainers can drive the MLers back to their
> ML, leaving a flame/dead group (for which the regulars here bemoan the
> lack of a way of cleaning up - that story's complicated).
You just said that people who opt out of mailing lists have reason for
complaining, but now you're saying that the complaining will drive them
back to the mailing lists. I can't even envision these conflicts you
propose, especially in this context. The lists and projects I've
mentioned demonstrate that there are many people, in many different
fields, who have productive and interesting things to say on this
newsgroup's topics, and I find the odds extremely slim that these ideas
will be drowned out by people fighting about who has a better mailing
list, or whether a mailing list is better than a newsgroup. But maybe
that's just me.
> Overall, what I'm saying is plug as many of the holes in your
> proposal you can to reduce the number of NO votes you might
> get from them. Did I mention anywhere that there are a fair
> number of reasons (not all reasonable) that some folks vote
> NO for? There's an FAQ on that, probably available at the
> same place you got the other creation FAQs ("Why People Vote
> NO", or something like that). Just remember that your
> readers will be voting YES, and that folks that think your
> group will damage their forum and news.group folks that
> think your proposal is flawed (poor charter, poor name,
> deceptive rationale) will vote NO. Thus, one of your jobs
> is actually to reduce the NO votes.
Thanks, I have already looked at the "Why people vote no" post. In
fact, we expected from the beginning that there could be political
issues here. Our approach has been to be very up front with the issues,
to recognize them and deal with them, but your comments sometimes make
me wonder if that was a mistake :-). Anyway, I'd understandably be more
concerned if I actually saw more widespread evidence here that there
were indeed problems.
-Dave
David C. DiNucci
> ... I can't imagine it. (Even in the
> bizarre chance that something like this happened, ...
Oops, I left out the smiley after considering the potential results of
an impossible hypothetical case. Suffice it to say the that goals for
the newsgroup are orthogonal to the goals of various groups' mailing
lists, but the same sorts of people will likely be interested in each.
-- Dave
ru.ig...@usask.ca
>So, in brief, you are willing to broaden the definition of "Usenet
>people" to those who can be shown, through a survey, to be willing to
>read or post to this newsgroup.
Not "broaden", this is historically what I've suggested
for proposals, especially of this type.
>I think that's exactly what a CFV is.
Not exactly what I am suggesting...
>So, you are recommending a straw vote?
That would be ok, but I think a survey would serve you better.
There's the matter of the "why"s, and other issues that would
fill in gaps in the proposal, that a survey can do but a
straw poll can't.
[snip]
>ru.ig...@usask.ca wrote:
>> You sort of implied the possibility when you say folks opt
>> out of the mailing lists for various reasons. They have
>> reasons for complaining, some of which is simply due to
>> traffic, others for the ML administration, etc. This implies
>> some degree of disfranchisement, which at higher degrees is
>> not desirable. We've had our share of those here.
[snip]
>I believe that few "folks opt out of mailing lists".
I was more or less alluding to the statement "many prospective
participants feel discouraged from subscribing" in the RFD.
It did say "many", not few.
[snip]
>I don't understand your contention that "they have reasons for
>complaining". Are you suggesting that, if people normally use a
>congested highway or eat at a restaurant with a strict dress code, and
>then find a less congested alternate route or a more casual restaurant,
>that that gives them reasons for complaining?
Of course it does. People complain about congested highways,
even after finding a less congested one (sometimes because
of it). People complain about restaurants with dress codes
(e.g. they want the food, but not have to dress up for it).
If the restaurant is the best in town, or the highway has
better road surfacing, or the mailing list has the best
people, folks that want in but can't for some reason are
going to complain. The proposal did say "moderators will
often discourage use of these lists [for less topical discussion]".
Some number of folks wouldn't be happy with that.
>And, moreover, that it is
>justification for closing down the alternate route or casual
>restaurant?
In this case, the alternate route and the restaurant don't
exist yet.
>On the other hand, given free reign, some discussion will
>tend to evaluate and discuss the pros and cons of the surrounding
>environment, and that can be a productive exercise.
>You are also the only one claiming "some degree of disenfranchisement".
Well, I DID say "implies some". After all, the proponents did
indicate that some folks go with some ML because the volume
was too high, one proponent indicated he was being asked not
to post so often, and there is that problem with fitting into
the topic space of the MLs. When folks get excluded like that
there's bound to be some sentiment of disfranchisement.
>Are you suggesting that people will drop out of the Grid Forum or
>Peer-to-Peer WG because there is a newsgroup in addition to their
>mailing lists?
In general, some fraction of folks do that, and that is often
the expectation for mailing lists (i.e. the list readers are
expected to move discussion totally out of the list and into
usenet).
>Or are you suggesting that people will move discussion
>and administration of the specific group projects off of those mailing
>lists and into a general newsgroup?
Administration, I wouldn't think so either. Moving some
discussion of projects out to newsgroups (and broaden the scope)
is what some groups are made for.
>I can't imagine it. (Even in the
>bizarre chance that something like this happened, if it helped the
>overall functioning of the Forum or WG, they would likely see it as a
>benefit!) It is also hard to think of someone feeling disenfranchised
>if a mailing list that was created (let's say) for a particular standard
>relating to p2p software issues doesn't want to accomodate more
>wide-ranging topics that might (for now) relate more to grids.
Well, it happens. Most haven't made it through the RFD phase.
Those are a pain in the rear. Excluding the soc.culture.*
perenial fiasco's (political disfranchisement is a mainstay of
our culture), the last bad one I remember is
soc.support.stalking-victims... That one looked ok on paper, but
it had a messy backstory.
>> >Nor do I understand what complainers would
>> >hope to gain, especially since people posting to the newsgroup will
>> >often likely be the same people (or work with people) posting to the
>> >lists.
>>
>> Hypothetically, the complainers can drive the MLers back to their
>> ML, leaving a flame/dead group (for which the regulars here bemoan the
>> lack of a way of cleaning up - that story's complicated).
>You just said that people who opt out of mailing lists have reason for
>complaining, but now you're saying that the complaining will drive them
>back to the mailing lists.
No, I'm saying the ones making the newsgroup into a sewer would
be the ones that had earlier opted out of the ML, then the current
MLers, who try to start using the new group find it to be a sewer
and stop using the newsgroup. The latter has happened, to the
extent that we usually encourage MLers to try out newsgroups
before attempting the the rest of the process here. Am I incorrect
in assuming the MLers in this case are familiar with newsgroups?
(once again, ru stumbles on a question he should have asked at
the beginning - *sigh* there's only so many issues I can remember
to ask about)
>The lists and projects I've
>mentioned demonstrate that there are many people, in many different
>fields, who have productive and interesting things to say on this
>newsgroup's topics,
Yabbut none of that matters, especially if you can't connect it
to usenet. Most of the Big-8 is used for non-productive,
non-interesting (to most Big-8 readers) things to say, and it's
all from folks who are trying to use the Big-8.
>and I find the odds extremely slim that these ideas
>will be drowned out by people fighting about who has a better mailing
>list, or whether a mailing list is better than a newsgroup. But maybe
>that's just me.
While, that's not what I had in mind (I was thinking more of
personal or project related stuff), when it comes to usenet,
I don't dismiss any kind of behaviour.
David C. DiNucci
>
> David C. DiNucci <da...@elepar.com> wrote:
> >I think that's exactly what a CFV is.
>
> Not exactly what I am suggesting...
Whether it is what you were suggesting or not, that is what a CFV is.
> >I don't understand your contention that "they have reasons for
> >complaining". Are you suggesting that, if people normally use a
> >congested highway or eat at a restaurant with a strict dress code, and
> >then find a less congested alternate route or a more casual restaurant,
> >that that gives them reasons for complaining?
>
> Of course it does. People complain about congested highways,
> even after ...
You first claimed that providing alternatives gives reasons for
complaining, and repeat that now with "of course it does", but then
explain it all by saying that providing alternatives doesn't always stop
people from complaining about things that predated the existence of the
alternatives. I've wasted too much of my time, and potentially other
readers' time, dealing with comprehension problems. If you still have
concerns, I'll live with it.
> >Are you suggesting that people will drop out of the Grid Forum or
> >Peer-to-Peer WG because there is a newsgroup in addition to their
> >mailing lists?
>
> In general, some fraction of folks do that, and that is often
> the expectation for mailing lists (i.e. the list readers are
> expected to move discussion totally out of the list and into
> usenet).
This just further illustrates that this discussion is a waste of time.
These groups are not just mailing lists, they are collections of
extremely smart people, many from top institutions around the world, who
work hard together to solve extremely difficult problems and attain
cutting-edge goals. The presence or absence of a newsgroup is certainly
not going to deter them. It would be a nice resource for some of us,
though.
> >and I find the odds extremely slim that these ideas
> >will be drowned out by people fighting about who has a better mailing
> >list, or whether a mailing list is better than a newsgroup. But maybe
> >that's just me.
>
> While, that's not what I had in mind (I was thinking more of
> personal or project related stuff), when it comes to usenet,
> I don't dismiss any kind of behaviour.
And I assure you, if anything turns off prospective voters, it will be
the fear that discussions such as the one we just had will take place on
this new newsgroup. I am confident, after observing other open venues
where we communicate, that that will not happen.
We have proposed a newsgroup, to serve a purpose which we have made very
clear. If people want this group, they will vote for it. If they
don't, they will vote against it. End of story. If the votes don't
materialize, for any reason, oh well. If anyone has any serious issues
or questions before the vote, I will be happy to address them.
-- Dave
David C. DiNucci
>
> David C. DiNucci <da...@elepar.com> wrote:
>
> >(b) Of course, the G5 problem is waiting to bite us again on the CFV
> >(and the followup RFD, if one is required), and I don't know how to stop
> >it. I understand that the CFV is usually posted to the same groups as
> >the RFD, and in fact, we would like it to be posted to the same groups.
> >Since we see the problems up ahead, does anybody have suggestions on how
> >to steer around them?
>
> I think what has been recommended up to now is to trim down the
> distribution, and post NEUTRAL pointers to "the very low traffic"
> news.announce.newgroups in the remaining groups. Not sure about
> the timing, or if the n.a.n moderator handles that. Consult the
> mentors on the wording of the pointer as well as the procedure.
> This might not be an issue if you get no one in those groups telling
> you they didn't see the RFD (in response to your followups).
Thanks much, that sounds like it will work fine for this case.
-- Dave
David C. DiNucci
>
> In article <3BDD9BAA...@elepar.com>,
> David C. DiNucci <da...@elepar.com> wrote:
>
> >(a) I see that a CFV cannot occur within 10 days of the last RFD, so if
> >we need another RFD, we should make it soon. So far, the only
> >suggestions I have seen here have been for very minor wording changes,
> >re-ordering of words, omitting words, etc., and I don't know if these
> >justify another RFD, but if they do, I'd like to know now if possible.
>
> No, they don't. You can make the changes on the final charter you
> send to be put on the CFV. That's the one that ends up being recorded
> everywhere.
Good to hear.
> >(b) Of course, the G5 problem is waiting to bite us again on the CFV
> >(and the followup RFD, if one is required), and I don't know how to stop
> >it. I understand that the CFV is usually posted to the same groups as
> >the RFD, and in fact, we would like it to be posted to the same groups.
> >Since we see the problems up ahead, does anybody have suggestions on how
> >to steer around them?
>
> Yeah, you might be able to convince the UVV person into multi-posting
> the CFV.
It does sound like maybe we should be discussing this with UVV. Thanks.
-- Dave
ru.ig...@usask.ca
>alternatives. I've wasted too much of my time, and potentially other
>readers' time, dealing with comprehension problems. If you still have
>concerns, I'll live with it.
Sorry about that. Sometimes I go too far or the wrong way
to try to settle a niggle at the back of my mind. Unfortunately
for you, the niggling this time just seems to be worse than with
other proposals.
David C. DiNucci
> ...
> Sorry about that. Sometimes I go too far or the wrong way
> to try to settle a niggle at the back of my mind. Unfortunately
> for you, the niggling this time just seems to be worse than with
> other proposals.
And I perhaps went too far when by predicting what kind of discussion
will not happen on the new group, but in general, these people have much
too much work to do to spend much time bickering. Productive debates
(leading to an occasional "aha!") could be another matter.
Thanks again,
Eugene Miya
This is meant to be a warning:
The proposal does not present an adequate justification for making
the group unmoderated. It needs to address this.
My personal general policy is to either vote Yes for Any proposal
or abstain. I very rarely vote No. And I rarely abstain.
I think that thos people who make poor news group proposals deserve
what they get. John Moore and the creation of sci.environment was
merely one example of this. I felt bad that Joh only came to that
realization too late.
I just took over the moderation of comp.parallel.
I can now see what my predecessor had to deal with with spam.
It overwhelmed him. He wasn't a good enough programmer to combat the problem.
I doubt that any one is yet capable of coding something that good.
Yet.
So the point of the warning is:
do you really want to do this, this way?
If you want to be smart, do something like Peter Neumann and comp.risks
or some of the moderated sci.* groups and get professional society
backing and moderate the group for money. Forget those who argue that
the delay is too long. They haven't been on the net long enough to know
what a real delay is.
You practically have to assume that you will get approval.
So set something up smart.
On the issue of news group creation:
In the past groups got created because traffic volume.
There is a newbie fantasy that creating a news group creates topic
oriented traffic. It's just that a fantasy. Sometimes it works, and
sometimes, it doesn't. This is independent of spammers.
--eugene miya, NASA Ames Research Center, eug...@ames.arc.nasa.gov.ha
Resident Cynic, Rock of Ages Home for Retired Hackers
Now Moderator comp.parallel
My 2nd favorite use of a flame thrower is Carpenter's remake of "The Thing."
A Ref: Uncommon Sense, Alan Cromer, Oxford Univ. Press, 1993.
David C. DiNucci
>
Thanks for the comments/advice, Eugene. (And thanks for getting the RFD
through comp.parallel moderation so quickly.)
> I have no basic problems with the proposed creation except one.
> This is meant to be a warning:
> The proposal does not present an adequate justification for making
> the group unmoderated. It needs to address this.
Well, I wouldn't be a proponent (and might not even vote) if it was
moderated. I personally got involved as a proponent here because:
* I would like a place to read and carry on free-form discussion on the
topics chartered,
* I have tried to carry on discussions on moderated newsgroups with very
limited success...well, let's face it, total failure
* I have tried to stretch the natural limits of certain mailing lists to
carry on more general discussions there, and have been advised against,
* I have seen quite a number of productive technical unmoderated
newsgroups, and
* I have seen productive unmoderated mailing lists, targeted
specifically to subsets of comp.distributed topics
> My personal general policy is to either vote Yes for Any proposal
> or abstain. I very rarely vote No. And I rarely abstain.
So I guess you vote Yes a lot!
> I think that thos people who make poor news group proposals deserve
> what they get. John Moore and the creation of sci.environment was
> merely one example of this. I felt bad that Joh only came to that
> realization too late.
You've tempted me to look at sci.environment. Most posts I see there do
indeed concern the environment, and the volume there looks pretty good
(e.g. one or two dozen per day), so I'll guess that the problem is that
many don't concern "sci", and therefore don't match the charter (and
John's intentions). In fact, it appears that many articles are x-posted
to talk.environment, where they are probably entirely on-topic. (I
personally DON'T think that moderation would be a good solution to that,
though I can imagine some sort of despam-bot or newsreader filter that
circumvented such x-posted posts.)
Regarding comp.distributed, I (and maybe Raj, I don't know) recognize
that some social issues are part of the natural landscape for us.
Remember, the term P2P came into most people's lexicon with Napster,
which started as an attempt to circumvent copyright laws, and then
became sort of a symbol of freedom from central authority. We are not
going to pretend that these factors won't spawn discussion, or that they
can always be cleanly separated from the more technical issues. In
fact, our charter explicitly includes discussion of decentralized
authority and intellectual property issues, as they fit into this larger
picture of distributed resource collectives.
Having said that: Newsreaders are becoming more useful every day, and
if others read Usenet like I do, they focus on a few threads that
interest them at any one time, and ignore the rest until something
catches their eye. I notice that comp.software.patterns recently went a
step further and recommended thread prefixes as part of the charter. I
would personally prefer that such measures be utilized less formally,
if/when needed, and that group splits occur when certain topics become
busy and cohesive enough.
I would highly recommend that interested parties take a peek at
http://groups.yahoo.com/group/decentralization/messages/
What you will see is current projects, ideas, lots of technical
discussion, and yes, some politics, regarding P2P, all mixed together
and feeding off of each other. What you won't see, though, is more
discussion relating to grid issues, or a decentralized Usenet-like
infrastructure, or a very flexible interface for skipping through the
messages and/or managing threads. I think that group has pretty much
maxed out its current home.
> I just took over the moderation of comp.parallel.
> I can now see what my predecessor had to deal with with spam.
> It overwhelmed him. He wasn't a good enough programmer to combat the problem.
> I doubt that any one is yet capable of coding something that good.
> Yet.
I would be very interested in knowing what you consider spam there, what
you consider too much, and why you think it is being sent to
comp.parallel (assuming it has nothing to do with parallel processing).
> So the point of the warning is:
> do you really want to do this, this way?
>
> If you want to be smart, do something like Peter Neumann and comp.risks
> or some of the moderated sci.* groups and get professional society
> backing and moderate the group for money. Forget those who argue that
> the delay is too long. They haven't been on the net long enough to know
> what a real delay is.
I have been around since the early 80's, and I'm one of those arguing,
having tried to carry on discussions on moderated groups a few times.
It gets very tiring, waiting for at least two passes of moderation (for
both the original post and a reply) for each of my posts. On some
groups, these cycles can take days, and on others (e.g.
comp.lang.visual), I've had some posts just never show up, terminating
the discussion for reasons completely unknown and uncontrolled to the
discussants. In each case, we just moved to (or you might say
"infiltrated" or "homesteaded") an unmoderated group (e.g.
comp.lang.visual.prograph, in the last case), and everything worked out
fine.
In fact, speaking of comp.parallel, I know that you have often
"self-moderated" your posts for quite awhile. And even with your new
improved moderation, I still see some general parallel discussion taking
place on comp.parallel.pvm, for example, instead of comp.parallel, and I
am sure it is because of the presence of moderation. I also look at
unmoderated groups like comp.arch and comp.sys.super, and I don't see
much spam (perhaps in part by their unique way of handling off-topic
posts like homework questions).
Besides, we won't be a real group until Mentifex shows up! :-)
> You practically have to assume that you will get approval.
> So set something up smart.
>
> On the issue of news group creation:
>
> In the past groups got created because traffic volume.
>
> There is a newbie fantasy that creating a news group creates topic
> oriented traffic. It's just that a fantasy. Sometimes it works, and
> sometimes, it doesn't. This is independent of spammers.
I admit that this is my biggest question mark. I haven't any idea how
many people could even be reading this discussion. After seeing how
quickly the Grid Forum and the P2P WG blossomed, and seeing the message
traffic elsewhere on related topics, I'm sure that there are plenty of
people who have many interesting ideas, but I may not know until the CFV
how many people are really considering Usenet as a good place to do
that. I'm not much of a campaigner, I'm just facilitating the
opportunity for others to become involved in creating a group that I,
myself, would like to exist.
> Resident Cynic, Rock of Ages Home for Retired Hackers
(Some day I'll qualify for a t-shirt!)
Thanks for the feedback,
-Dave
Eugene Miya
David C. DiNucci <da...@elepar.com> wrote:
>Thanks for the comments/advice, Eugene. (And thanks for getting the RFD
>through comp.parallel moderation so quickly.)
n.a.n. approved it. I didn't see it. Dave Lawrence (tale)) or one of
his successors approved it as apart of a cross post. I would have
approved it had I seen it as per long time Usenet tradition, but c.p.
only has cross post agreements with a smaller number of groups.
The news.* hierarchy of course has a different status. I think that
some groups should be x-posted to news.groups, but I draw a line at my
net involvement. Things are "different" at work, since you left.
>Eugene Miya wrote:
>> The proposal does not present an adequate justification for making
>> the group unmoderated. It needs to address this.
>
>Well, I wouldn't be a proponent (and might not even vote) if it was
>moderated. I personally got involved as a proponent here because:
Note: I said the proposal. Discussion can bring this out, and all of
your arguments as you known have frequently been presented before
(big surprise of course).
>* I have tried to carry on discussions on moderated newsgroups with very
>limited success...well, let's face it, total failure
No, I think Steve has c.p. going and Peter does pretty well in comp.risks.
>* mailing lists
Don't scale.
>> My personal general policy is to either vote Yes for Any proposal
>> or abstain. I very rarely vote No. And I rarely abstain.
>
>So I guess you vote Yes a lot!
Used to, I could be counted upon as a easy vote.
Rememer: poor proposals deserve what they get.
>> I think that thos people who make poor news group proposals deserve
>> what they get. John Moore and the creation of sci.environment was
>> merely one example of this. I felt bad that Joh only came to that
>> realization too late.
>
>You've tempted me to look at sci.environment. Most posts I see there do
>indeed concern the environment, and the volume there looks pretty good
>(e.g. one or two dozen per day), so I'll guess that the problem is that
>many don't concern "sci", and therefore don't match the charter (and
>John's intentions). In fact, it appears that many articles are x-posted
>to talk.environment, where they are probably entirely on-topic. (I
>personally DON'T think that moderation would be a good solution to that,
>though I can imagine some sort of despam-bot or newsreader filter that
>circumvented such x-posted posts.)
Well at the time the talk.* hierarchy had greater plans. It was
intended to be quite respectable until bureaucrats and others started to
throw it out (including Milo). We clearly have different ideas on
moderated/unmoderated combinations these days.
>Regarding comp.distributed, ...
...
>Remember, the term P2P came into most people's lexicon with Napster,
>which started as an attempt to circumvent copyright laws, and then
>became sort of a symbol of freedom from central authority. ...
>...
I am not at liberty to discuss certain P2P issues.
>thread prefixes
Been around a while with mixed success.
Not at liberty to discuss them either.
>I would be very interested in knowing what you consider spam there, what
>you consider too much, and why you think it is being sent to
>comp.parallel (assuming it has nothing to do with parallel processing).
The messages coming in which aren't aligned with the group (as Mike
informed me) include the whole gamet: sex web sites, Ponzi posts,
mail list DB sites. And so forth. Stuff having to do little with
parallel computing. One did attempt to advertise for a search engine
(marginally parallel).
I suspect that there is a wavefront of spam posts which reach some
groups and not others. It depends on how an active file is configured
and how much script and header editing the spammer does. A few are
smarter attempt to randomize the groups they hit.
The usual stuff argued for mod/unmod snipped.
>Besides, we won't be a real group until Mentifex shows up! :-)
Arthur is an amusing human.
Naw he has a ways to go. He reads too much sci-fi, and he hasn't read
enough material critical of AI.
>> Resident Cynic, Rock of Ages Home for Retired Hackers
>
>(Some day I'll qualify for a t-shirt!)
Well, saves on postage if you qualify during JCDL Program Committee
meetings in Portland.
David C. DiNucci
>
> In article <3BDF80D4...@elepar.com>,
> David C. DiNucci <da...@elepar.com> wrote:
> >Thanks for the comments/advice, Eugene. (And thanks for getting the RFD
> >through comp.parallel moderation so quickly.)
>
Well, thanks to tale then.
> Things are "different" at work, since you left.
Gee, thanks. :-) (I even read that Dan himself is resigning.)
> >Eugene Miya wrote:
> >Well, I wouldn't be a proponent (and might not even vote) if it was
> >moderated. I personally got involved as a proponent here because:
>
> Note: I said the proposal. Discussion can bring this out, and all of
> your arguments as you known have frequently been presented before
> (big surprise of course).
Right, no surprise, and maybe that's why unmoderated groups are
commonplace. Raj may or may not agree with my reasons exactly, but he
and Mark Baker proposed it as unmoderated before I even got involved.
If this group can't facilitate free discussion, I believe it's
worthless. I tolerate some level of spam every day.
> >* I have tried to carry on discussions on moderated newsgroups with very
> >limited success...well, let's face it, total failure
>
> No, I think Steve has c.p. going and Peter does pretty well in comp.risks.
I thought I read that you now had c.p, but regardless: OK, so some
discussions might actually be reasonable on a moderated group, if the
moderator for that group at the time happens to be very devoted to it,
and doesn't happen to be unavailable when the discussion is taking
place, and doesn't have some personal objection to the topic, etc.,
etc., but once these constraints aren't met a few times, disenchantment
sets in quickly. I mentioned earlier that the "decentralization" list
sometimes gets many messages/hour (of back-and-forth discussion), and
someone even responded here that that was hardly enough to justify a
newsgroup, but forget about moderating that. What is too few for some
is too many for others.
> >* mailing lists
>
> Don't scale.
Many different interpretations. If you mean they don't scale up well,
in that they have limited distribution and/or browsing infrastructure, I
agree.
> Well at the time the talk.* hierarchy had greater plans. It was
> intended to be quite respectable until bureaucrats and others started to
> throw it out (including Milo).
All the more explanation of why the t.e posts would be x-posted to s.e I
guess, if the latter is the only "respectable" place where they are even
remotely on-topic.
> I am not at liberty to discuss certain P2P issues.
I always expected we'd have lurkers.
> >http://groups.yahoo.com
>
> Not at liberty to discuss them either.
I don't get that one. (And, I guess I won't.)
> >I would be very interested in knowing what you consider spam there, what
> >you consider too much, and why you think it is being sent to
> >comp.parallel (assuming it has nothing to do with parallel processing).
>
> The messages coming in which aren't aligned with the group (as Mike
> informed me) include the whole gamet: sex web sites, Ponzi posts,
> mail list DB sites. And so forth. Stuff having to do little with
> parallel computing. One did attempt to advertise for a search engine
> (marginally parallel).
And those posts were so prevalent that it was tough for the moderator to
keep up? Wow. I still wonder why I don't see them in unmoderated
groups.
> I suspect that there is a wavefront of spam posts which reach some
> groups and not others. It depends on how an active file is configured
> and how much script and header editing the spammer does. A few are
> smarter attempt to randomize the groups they hit.
If this is a justification for moderation, it would seem to apply to
every group on Usenet, and moderating them all has not only proven
unnecessary, it could make Usenet as clogged as our airports these
days. I vastly prefer to go with the lightweight solution, and add
overhead as circumstances warrant.
> Well, saves on postage if you qualify during JCDL Program Committee
> meetings in Portland.
I'm also signed up for the IPG meeting in Dec.
Thanks,
-- Dave
Rajkumar Buyya
"David C. DiNucci" wrote:
>
> > >Eugene Miya wrote:
>
> > >Well, I wouldn't be a proponent (and might not even vote) if it was
> > >moderated. I personally got involved as a proponent here because:
> >
> > Note: I said the proposal. Discussion can bring this out, and all of
> > your arguments as you known have frequently been presented before
> > (big surprise of course).
>
> Right, no surprise, and maybe that's why unmoderated groups are
> commonplace. Raj may or may not agree with my reasons exactly, but he
> and Mark Baker proposed it as unmoderated before I even got involved.
> If this group can't facilitate free discussion, I believe it's
> worthless. I tolerate some level of spam every day.
I generally have greater degree of tolerance for all those messages that are
related at least to technical aspects of computing or something that open up new
opportunities, etc. Only thing that I feel painful to deal with is: receiving
tens of the same message from the same person.
When I am on high volume mailing lists, I go for Digest Version, that helps
me to keep track of what is happening on the earth. But when I find
any interesting message, it is very hard for me to follow-up or be part
of that discussion with digest version.
Thanks Dave for answering all comments etc. raised on our proposal. I appreciate
you for taking lead on this. Well done!
cheers
Raj
Eugene Miya
>> No, I think Steve has c.p. going
>
Oops, that was supposed to be "had".
There are or were (changed over time) many fine, traffic-y, moderated
groups. Darrell Long had comp.os.research (it's a bit spotty now),
sci.military(.now.moderated in name, too), and lots of other groups.
Naw the real problem Dave is attention span.
But you already knew that.
>> >* mailing lists
>> Don't scale.
>
>Many different interpretations. If you mean they don't scale up well,
>in that they have limited distribution and/or browsing infrastructure,
>I agree.
Be glad that you aren't experiencing the big email lists at your former work.
It's especially amusing to those with last names toward the end of the list.
>> Well at the time the talk.* hierarchy had greater plans. It was
>> intended to be quite respectable until bureaucrats and others started to
>> throw it out (including Milo).
>
>All the more explanation of why the t.e posts would be x-posted to s.e I
>guess, if the latter is the only "respectable" place where they are even
>remotely on-topic.
Well the original theory back then was moderated discussion and
announcements could go out the sci.* group and general unmoderated talk
could happen there. And you see these attempts still going on in
subhierarchies.
No one ever said balancing "free speech" and filtering for content would
be easy. I'm impressed with the various net mechanisms which our
"forefathers" like Erik, Brian, and others put into place (like rot13
for instance to keep humor alive).
>> Not at liberty to discuss them either.
>I don't get that one. (And, I guess I won't.)
No, they have to do directly with work.
>And those posts were so prevalent that it was tough for the moderator to
>keep up? Wow. I still wonder why I don't see them in unmoderated
>groups.
They clobbered Mike. I just had to respond to RoadRunner (ISP) about
3 messages before I got to yours here.
>If this is a justification for moderation, it would seem to apply to
>every group on Usenet, and moderating them all has not only proven
>unnecessary, it could make Usenet as clogged as our airports these
>days. I vastly prefer to go with the lightweight solution, and add
>overhead as circumstances warrant.
That is an valid opinion. And you will see some, but I can see also
that other moderators are whacking at spammers before me as a further
line of defense to the unmoderated groups.
Just killing time between OCR scans Dave. Not clear how much longer
I will scan n.g.
David C. DiNucci
> There are or were (changed over time) many fine, traffic-y, moderated
> groups. Darrell Long had comp.os.research (it's a bit spotty now),
> sci.military(.now.moderated in name, too), and lots of other groups.
Well, I do hope discussion on moderated groups is feasible in case we
ever need to go in that direction, but for now, I vote for simplicity.
> Naw the real problem Dave is attention span.
> But you already knew that.
Two sides of the same coin. Good discussions often require attention,
and maintaining attention (either as a poster or a lurker) can be taxing
when it is stretched over days between messages while other
responsibilities impinge.
> Be glad that you aren't experiencing the big email lists at your former work.
> It's especially amusing to those with last names toward the end of the list.
Well, if they have anything to do with grids (like some that I set up
before I left), I guess members should vote for this newsgroup. :-) (In
fact, I recall I was looking into creating local newsgroups even back
then, but propogation to all interested parties seemed iffy.)
> No one ever said balancing "free speech" and filtering for content would
> be easy. I'm impressed with the various net mechanisms which our
> "forefathers" like Erik, Brian, and others put into place (like rot13
> for instance to keep humor alive).
Agreed.
> Just killing time between OCR scans Dave. Not clear how much longer
> I will scan n.g.
Thanks for the feedback you've given. Every little bit helps.
Jay Denebeim
>I suspect that there is a wavefront of spam posts which reach some
>groups and not others. It depends on how an active file is
>configured and how much script and header editing the spammer does.
>A few are smarter attempt to randomize the groups they hit.
I suppose. I don't see very much spam at all though. Almost none in
fact. Of course, I, and all my peers use cleanfeed or spamhippo.
Those do a great job of getting rid of the spam that gets posted to
usenet. That leaves just the spam coming to the moderation queues,
these are taken care of by registration to the newsgroup. Spammers
always mung their address, so if you require a registration for the
newsgroup the spam goes to 0.
Greg Pfister
>
> David C. DiNucci <da...@elepar.com> wrote:
>
> >I admit to being a bit confused by what constitutes acceptable evidence
> >of Usenet popularity.
>
> Basically, either a) a lot of folks already posting to usenet on
> the topic, b) a lot of folks who say they are reading usenet
> postings on the topics, c) a lot of folks not on usenet who say
> they want to read a newsgroup on the topic.
How about d) a lot of folks already posting to usenet who would
post on this topic, but don't have a place to do so?
Like me. Like a number of others who post to comp.arch,
comp.sys.super, and the like, who being decent netizens avoid
posting OT questions like "what's the difference between P2P and
Grid, anyway?" to those groups, although I can tell from context
that they would *definitely* have opinions.
(And such questions are, by the way, guaranteed to (a) arise; (b)
produce volumes and volumes of opinion. The TFCC mailing list was
inundated when "what is a cluster, anyway?" was discussed. The
others are even more subject to opinion.)
--
Greg Pfister http://pfister.userv.ibm.com/
Eugene Miya
you can set up an alt.* group. Forget the voting. There are alt.sci.*
groups and one could just as easily set up alt.comp.* groups.
Gotta run.
David C. DiNucci
Thanks, Greg. And, I would add, that these general discussions are more
than fluff. They shed light on *why* there are different approaches,
which differences are important as opposed to purely historical, and how
these concepts are related to existing ones.
After seeing the "Gathering Traffic Data for Proposed Newsgroups" post
the other day, I decided to see what I could find out about the existing
use of Usenet for P2P and Grid discussion, using groups.google. It turns
out to be a very difficult question to answer in this case, because of
the "overloaded" terms.
For example, p2p can stand for everything from peer-to-peer to
person-to-person (or people-to-people) to "parent to parent" to
"positive to positive" to phenylacetone (phenyl-2-propanone, an
amphetamine). Even when it stands for "peer-to-peer", it can also be a
technical networking term (as opposed to client-server) or a description
of people getting together (e.g. I know groups that have had regular
"Peer-to-Peer" get-togethers). And even when it means what we want it to
mean, it has often been used as an abbreviation of "Napster and other
similar ways of sharing music", so the focus of the article may be music
instead of p2p.
The term "grid" is even worse--everything from street grids to
electrical power grids (or even within a tube, "grid voltage"), and
mathematical (structured and unstructured) grids for simulation to
maps/Irish grids. It's even harder to separate these terms when you
realize that electrical grids are often used as an analogy to "our"
grids (e.g. NASA calls its grid "Information Power Grid"), and that
"our" grids are often used to run mathematical grid simulations.
Yet, almost everyone uses these terms unqualified, because their meaning
is "understood" from the context.
On top of that, google can return amazingly different numbers, depending
upon whether you accept its count (which includes all "similar" posts),
or actually look to the last page (which omits the similar posts). It
also doesn't like "to", and can't handle more than 10 words at a time in
its queries.
In that light, after lots of attempts to weed out false positives, here
are some google results of instances, in english, over the 3-month
period Aug 2 to Nov 2.
Query Number of hits
---------- ------------------
"distributed"
and either "grid" or "p2p"
but not "power", "map",
"uniform", "generation",
"utilities", "cells", or "control":
730 (or 3940?)
"information power grid": 21 (or 31)
both "p2p" and "grid"
but not "distributed": 39 (54)
p2p but not "distributed",
"amphetamine", "drugs",
"music", "mp3",
or "person-to-person": 855 (15,100?)
(Note that none of these three overlap with the first query.)
"network computing": 779 (or 25,400?)
"computational grid": 119
"computational grids": 164
"ian foster" and "grid":
but not "promo": 120
(Ian foster is a leading proponent of "computational" grids, but there
is also a musician named that, hence the "- promo")
"peer-to-peer": 890 (174,000?)
Although tough to pinpoint, I would conclude that there are at least
sufficient messages/day on this topic already on usenet, not even
considering the many mailing lists on the subject.
David C. DiNucci
I don't think the remaining time to create the group is a major factor
in the long run. (In fact, we apparently could already have initiated
the CFV procedure, but are waiting until next week to make sure
everybody gets the word first.) Overall, I think the improved
distribution and awareness level of a big-8 group will be well worth the
wait.
ru.ig...@usask.ca
>Greg Pfister wrote:
>>
>> ru.ig...@usask.ca wrote:
>> >
>> > David C. DiNucci <da...@elepar.com> wrote:
>> >
>> > >I admit to being a bit confused by what constitutes acceptable evidence
>> > >of Usenet popularity.
>> >
>> > Basically, either a) a lot of folks already posting to usenet on
>> > the topic, b) a lot of folks who say they are reading usenet
>> > postings on the topics, c) a lot of folks not on usenet who say
>> > they want to read a newsgroup on the topic.
>>
>> How about d) a lot of folks already posting to usenet who would
>> post on this topic, but don't have a place to do so?
I left that out intentionally. In many cases, if the topic
is popular enough, it does find a place where it is either
marginally on-topic, is off-topic but tolerated, or is
completely on-topic but within a generalized forum. It
might not be the bulk of the readership, but it should be
there if the readership is large. There simply aren't
many topics in usenet that aren't represented somehow
in a newsgroup, even as a minority topic.
Regardless, a), b), and d) are preferable to mailing
list or web board statistics mainly because of their direct
association with usenet.
>Thanks, Greg. And, I would add, that these general discussions are more
>than fluff. They shed light on *why* there are different approaches,
>which differences are important as opposed to purely historical, and how
>these concepts are related to existing ones.
No one really cares if the discussion is "fluff" or "shed light".
Frankly, we are totally willing to accept fluff groups over an
important group. Remember my rant:
"Quality, usefulness, merit, or non-newsgroups popularity of a topic
is more or less irrelevant in creating a new Big-8 newsgroup.
Usenet popularity is the primary consideration.
Fluff is ok. All that matters is if you have enough readers.
[snip]
>On top of that, google can return amazingly different numbers, depending
>upon whether you accept its count (which includes all "similar" posts),
>or actually look to the last page (which omits the similar posts).
The archive searches have historically had that problem. I thought
that was mentioned in the FAQ, and that the only numbers that have
a chance of meaning something is the "last page" stats. Keep in
mind that the part that says "Results n1 - n2" is the more like
the number of THREADS it finds and the "about n3" is the number of
messages. When you don't have the "about" then you are stuck with
trying to figure how many messages that actually means (brute force
counting?).
>It
>also doesn't like "to", and can't handle more than 10 words at a time in
>its queries.
>In that light, after lots of attempts to weed out false positives, here
>are some google results of instances, in english, over the 3-month
>period Aug 2 to Nov 2.
>Query Number of hits
>---------- ------------------
>"distributed"
>and either "grid" or "p2p"
>but not "power", "map",
>"uniform", "generation",
>"utilities", "cells", or "control":
> 730 (or 3940?)
I think you made a mistake. I only get 273 hits. Did you remember
to click on "Return messages posted between" radio button? If you
don't do that you get the stats you mention, which cover 1995 to
now. If you select that properly, you get 273 hits ...which, by
the way includes the RFD discussion (news.groups stuff should be
excluded from the count). After excluding those, you still have
(using sort by date), a lot of totally off topic postings (for the
reasons you stated). I always recommend sifting through a hundred
or so of the postings to get a rough estimate of the fraction of
on-/off-topic stuff. Keep in mind that the links displayed often
represent more than one posting BUT the Google coders didn't
see fit to provide a way to indicate which ones went into the
hit stats.
>p2p but not "distributed",
>"amphetamine", "drugs",
>"music", "mp3",
>or "person-to-person": 855 (15,100?)
>(Note that none of these three overlap with the first query.)
Same mistake, and same problems. I've forgotten how to exclude
newsgroups from a search. Need to trash those jobs postings, too.
"-group:*job*"? Alt.adoption seems to be a huge contributer, too.
>Although tough to pinpoint, I would conclude that there are at least
>sufficient messages/day on this topic already on usenet, not even
>considering the many mailing lists on the subject.
There looks to upwards of 50% off-topic traffic in the properly
assigned time frame. Between that and the error in the search,
that conclusion will have to wait a bit.
David C. DiNucci
> >Thanks, Greg. And, I would add, that these general discussions are more
> >than fluff. They shed light on *why* there are different approaches,
> >which differences are important as opposed to purely historical, and how
> >these concepts are related to existing ones.
>
> No one really cares if the discussion is "fluff" or "shed light".
You are over-generalizing. There may be many people that care
(including me, for that matter), maybe even care enough to decide not to
vote for the group if it will contain mostly fluff. You are welcome not
to care.
> Frankly, we are totally willing to accept fluff groups over an
> important group. Remember my rant:
When you say "we are willing to accept", are you suggesting that there's
some well-defined group that decides whether to accept a group or not,
regardless of the vote resulting from the CFV? (It's a serious
question.) If yes, do you believe that you are a member of that group?
Please point me to some documentation.
> >On top of that, google can return amazingly different numbers, depending
> >upon whether you accept its count (which includes all "similar" posts),
> >or actually look to the last page (which omits the similar posts).
>
> The archive searches have historically had that problem. I thought
> that was mentioned in the FAQ, and that the only numbers that have
> a chance of meaning something is the "last page" stats.
The FAQ seems to imply that the "about" numbers it gives are just plain
wrong. However, it appears that they might be about right if one clicks
the link on the last page that requests that all similar posts be
included. In other words, it seems to represent ALL posts on the
subject, rather than just unique-looking ones.
> Keep in
> mind that the part that says "Results n1 - n2" is the more like
> the number of THREADS it finds and the "about n3" is the number of
> messages. When you don't have the "about" then you are stuck with
> trying to figure how many messages that actually means (brute force
> counting?).
Your understanding of the numbers matches mine (if you interpret
"similar" posts as belonging to the same thread), so I don't understand
your first statement, that "the only numbers that have a chance of
meaning something is the last page stats". You recognize that those
reflect only the number of threads and that we really want to know the
number of messages, which is present in the "about n3" number (or the
last page stats only if you hit the "include similar posts" link on the
last page).
Regardless, I have included both the thread and message counts. People
can come to their own conclusions.
> >In that light, after lots of attempts to weed out false positives, here
> >are some google results of instances, in english, over the 3-month
> >period Aug 2 to Nov 2.
>
> >Query Number of hits
> >---------- ------------------
> >"distributed"
> >and either "grid" or "p2p"
> >but not "power", "map",
> >"uniform", "generation",
> >"utilities", "cells", or "control":
> > 730 (or 3940?)
>
> I think you made a mistake. I only get 273 hits. Did you remember
> to click on "Return messages posted between" radio button?
Memory had nothing to do with it. User interface design and missing
instructions did. Yes, I came up with the same count as you after
hitting the radio button.
> If you
> don't do that you get the stats you mention, which cover 1995 to
> now. If you select that properly, you get 273 hits ...which, by
> the way includes the RFD discussion (news.groups stuff should be
> excluded from the count).
Exclude what you like.
> After excluding those, you still have
> (using sort by date), a lot of totally off topic postings (for the
> reasons you stated). I always recommend sifting through a hundred
> or so of the postings to get a rough estimate of the fraction of
> on-/off-topic stuff. Keep in mind that the links displayed often
> represent more than one posting BUT the Google coders didn't
> see fit to provide a way to indicate which ones went into the
> hit stats.
I have already explained the 10 word limit that keeps me from filtering
more. You are welcome to do whatever searches you like. I've given you
hints of where to start.
> >p2p but not "distributed",
> >"amphetamine", "drugs",
> >"music", "mp3",
> >or "person-to-person": 855 (15,100?)
> >(Note that none of these three overlap with the first query.)
>
> Same mistake, and same problems. I've forgotten how to exclude
> newsgroups from a search. Need to trash those jobs postings, too.
> "-group:*job*"? Alt.adoption seems to be a huge contributer, too.
I don't know about huge, but that's why I implied "p2p" might mean
"parent-to-parent".
> >Although tough to pinpoint, I would conclude that there are at least
> >sufficient messages/day on this topic already on usenet, not even
> >considering the many mailing lists on the subject.
>
> There looks to upwards of 50% off-topic traffic in the properly
> assigned time frame. Between that and the error in the search,
> that conclusion will have to wait a bit.
Feel free to filter them as you wish, and come to your own conclusions.
-Dave
ba...@dmcom.net
>
> ru.ig...@usask.ca wrote:
<snip>
> >
> > No one really cares if the discussion is "fluff" or "shed light".
>
> You are over-generalizing. There may be many people that care
> (including me, for that matter), maybe even care enough to decide not to
> vote for the group if it will contain mostly fluff. You are welcome not
> to care.
I do need to make a small comment, fluff does matter one way or another
to some users. Some have posted leaving moderated because of too mch
fluff, others post they enjoy the fluff, and some says fluff is better
then flames.
>
> > Frankly, we are totally willing to accept fluff groups over an
> > important group. Remember my rant:
>
> When you say "we are willing to accept", are you suggesting that there's
> some well-defined group that decides whether to accept a group or not,
> regardless of the vote resulting from the CFV? (It's a serious
> question.) If yes, do you believe that you are a member of that group?
> Please point me to some documentation.
No you misunderstand. The we refers to news admins and/or regs of
this group. Yes the CFV is all that matters for most NSPs (protests,
vote fruad, tech issues can cause exceptions)
<snip>
> Feel free to filter them as you wish, and come to your own conclusions.
As each that takes on the tasks shall do.
Comp.* in general have has good vote turn out and passed, but this
does not happen all the time. The concern about a valid search of
estimated existing posts (that you want in the group) is in a large part
a concern about a group with little or no traffic or some other
problems like a spam attractor.
In the end the Results matter, the question is raised though and now
often, why even go for a CFV using the resouces of people and networks
if only 10 people even discuss the topic and/or vote.
I am not saying that that only 10 people discuss what would be on-topic
for your proposal, in fact I am sure many more do in this case. Traffic
stidies (refined and filtered) and straw polls do serve as an indication
of how much interest there is. So you will see some requests to justify
that there is a need/interest for a group, that is your job as a
proponent.
--
news:alt.pagan FAQ at http://www.dmcom.net/bard/altpag.txt
news:alt.religion.wicca FAQ at http://www.dmcom.net/bard/arwfaq2.txt
news:news.groups (draft)FAQ at http://www.dmcom.net/bard/ngfaq.txt
ru.ig...@usask.ca
>ru.ig...@usask.ca wrote:
>>
>> David C. DiNucci <da...@elepar.com> wrote:
>> >Thanks, Greg. And, I would add, that these general discussions are more
>> >than fluff. They shed light on *why* there are different approaches,
>> >which differences are important as opposed to purely historical, and how
>> >these concepts are related to existing ones.
>>
>> No one really cares if the discussion is "fluff" or "shed light".
>You are over-generalizing. There may be many people that care
>(including me, for that matter), maybe even care enough to decide not to
>vote for the group if it will contain mostly fluff. You are welcome not
>to care.
>> Frankly, we are totally willing to accept fluff groups over an
>> important group. Remember my rant:
>When you say "we are willing to accept", are you suggesting that there's
>some well-defined group that decides whether to accept a group or not,
>regardless of the vote resulting from the CFV? (It's a serious
>question.) If yes, do you believe that you are a member of that group?
>Please point me to some documentation.
The creation FAQs basically indicate this. The group passes or
fails on the number of people who want it against the number
of people who USUALLY only have technical reasons for voting at
all. If the group doesn't get enough YES votes, it isn't accepted
no matter how important the topic is. If it gets too many NO votes,
it isn't accepted, no matter how important the topic is. Most
regulars beyond the usual 20 or so, usually do not vote at all
unless it is to vote NO due to a technicality. Even then, the
latter is contentious at times. Regardless, being a fluff group
is not considered a technicality, so news.group regulars are
willing to let a fluff group through. There are sometimes folks
interested in the topic that will vote NO because of some sort of
overriding concern, but that means they won't be voting on account
of it being a fluff group. Then there is the rest of Big-8 readers,
who simply don't care about what goes through, so the existence of
a fluff group is acceptable to them, too. It should be: a large
fraction of the groups, and certainly of the volume, are probably
considered fluff groups. Fluff gets more readers, so fluff has
a higher likelihood of passing, to the chagrin of folks who
struggle with groups for more important topics. So when I say
"we", I mean the Big-8 as an entity, or technically, the process,
is quite willing to accept a fluff group while tossing out more
important ones.
[snip]
>Your understanding of the numbers matches mine (if you interpret
>"similar" posts as belonging to the same thread), so I don't understand
>your first statement, that "the only numbers that have a chance of
>meaning something is the last page stats". You recognize that those
>reflect only the number of threads and that we really want to know the
>number of messages, which is present in the "about n3" number (or the
>last page stats only if you hit the "include similar posts" link on the
>last page).
Historically, the "about" used to fluctuate depending on how close
you got the the last page (it used to be a factor of 10 or 100
off). It used to use a guestimate algorithm (at a guess, I think
it extrapolated the total based on the first N hits by scaling
by the timeframe per page vs total - which sucked royally).
It's gotten better (new site, new code - it looks like they
actually count), but I still don't trust that about figure until
I get to the last page. Even then, the figure usually given looks
an awful lot like the bogus ones it used to give, but in the past
it used to enumerate each posting so we could see the last posting
number match the "about" (which turned into "exactly" or something
like that) figure. It doesn't do that anymore, so the proponent
has to struggle more with it.
David C. DiNucci
future than the past, while still remembering the past :-).
We are in final preparations of the Proponent Questionnaire for
comp.distributed. In an effort to progress smoothly toward a CFV, I want
to ensure that we have addressed the primary concerns raised in this
discussion on news.groups. Most recommendations were very clear, but
the last paragraph of the Rationale (which describes the relation of the
newsgroup to mailing lists, etc) prompted significant discussion. As a
result, I am proposing replacing that paragraph with:
----------Proposed new final paragraph of rationale -----------
Some existing newsgroups, like comp.parallel and comp.sys.super,
touch on specialized aspects of this topic, and will continue
to do so. However, this new group will allow more general
discussions, as well as serve as a focal point for considering the
inter-relationships, interactions, and synergies when combining
these separate technologies. In addition, several existing mailing
lists relate to comp.distributed topic areas (with sign-ups currently
totalling at least 650 and likely over 1000), but they generally focus
on discussion and administration of specific group projects and
priorities, and/or tend to segregate p2p and grid communities, even on
topics shared by those communities. The new newsgroup is expected to
complement these lists by providing a very broad venue for discussion
and interaction, while addressing the natural scaling problems which
would prevent these mailing lists from individually expanding their
charters to encompass all of the stated goals of this newsgroup.
---------------------------------------------------------------
I believe that this is, in general, a small enough modification not to
require another RFD. This, and deleting the "publicising" guidelines,
are the only changes I understand will be required to the rationale. In
later sections, I have made specific minor changes as suggested (below)
and have added the intention to post CFV pointers to alt.internet.p2p,
asu.comp.distributed, alt.gnutella, and alt.sci.seti. The remainder of
this message quotes related comments from Ru, as they relate to this and
other changes. If we missed something, I would appreciate it if someone
would bring it to our attention now.
Thanks,
-Dave
ru.ig...@usask.ca wrote:
>
> In news.groups Rajkumar Buyya <rajk...@csse.monash.edu.au> wrote:
> > REQUEST FOR DISCUSSION (RFD)
> > unmoderated group comp.distributed
...
> >Already, international academic and commercial forums like:
> [snip list]
> >have evolved to create standards and protocols for inter-operability
> >between heterogeneous systems providing virtual services. Recently,
> >infrastructure projects like the NSF Distributed TeraScale Facility
> >have focused even more attention, and include involvement from several
> >companies. Many computer and/or software vendors, large and small,
> >have recently announced specific projects or general priorities into
> >p2p and/or grids, including IBM, Intel, DSTC, Sun, and Microsoft.
> >Some details on these and other projects can be found at:
> [snip list]
>
> Ok, after this posting I'm definitely adding to my sig. Real world
> popularity and initiatives are of little concern here, and really
> have less impact on the success or failure of a proposal, generally
> speaking, than many proponents think. If the above vendors don't
> bother with newsgroups, their company's involvement is not going
> to mean a thing on the number of readers. If it is intended to
> impress readers, that's all it does, since really the only folks
> that would be impressed by the above are the ones already interested
> in the topic. It won't (shouldn't) move the folks that aren't
> interested in the topic because they aren't supposed to vote.
> I'd suggest throwing out the above paragraph.
>
> >Although over 20 discussion mailing lists operated by individuals or
> >institutions exist, they are generally intended for discussion of
>
> If you are going to cite a list, this is the one that might
> be of interest, along with readership stats of those who are
> interested in the new group. If you can't demonstrate these
> mailing lists actively support your initiative, there's no
> point in mentioning them here, either.
>
> >specific group priorities, and strongly segregate p2p and grid
> >communities, even when addressing similar issues. Another concern is
> >that mailing lists are likely to generate large volume of email for
> >members;
>
> Tell us how many is "large volume". No one should take that
> term seriously here without quantification.
>
> >therefore, moderators will often discourage use of these
> >lists for general or controversial discussion, and many prospective
> >participants feel discouraged from subscribing, do not become members,
> >and do not join important topical discussions. We believe that having
> >a newsgroup where people can participate in discussions of their own
> >choosing, when they want, without getting swamped with emails, will
> >help overcome these limitations and will encourage discussion and
> >dissemination without the need of explicit membership.
>
> Really, the excuse given above is always true for any mailing
> list, it's just a question of degree. I wouldn't mind seeing
> some supporting evidence that this is really a problem, e.g. a
> word from many of the mailing list maintainers. I'd really
> hate to see a potentially usable group be populated by
> malcontent rejects of mailing lists creating what should really
> be a .advocacy group, for example (though in that case, if
> there are that many malcontents the topic should have enough
> interested usenet readers, or be a total washout).
>
> New paragraph.
>
> >While some
> ^^^^^
> delete
> >existing newsgroups, like comp.parallel and comp.sys.super, touch on
> >some specialized aspects of this topic, and will continue to do so,
> >this new group will serve as a focal point for considering the inter-
> ^
> New sentence: "However, this new group will allow more general
> discussions, as well as serve as a focal..."
> >relationships, interactions, and synergies when combining these
> >separate technologies.
>
> And the above paragraph should come BEFORE your mailing list
> section.
Done.
> >Strategy for publicising the comp.distributed newsgroup:
>
> >The formation of the comp.distributed newsgroup will be publicised
> >through the following channels (but not limited to):
>
> [list snipped]
>
> I don't see the point in this list. Once the group is created,
> it out of the hands of us news.group folks, and the users of
> the new group are free not to advertise it. Frankly, if it
> passes, it should really be beyond the need for any hard selling.
> I'd say snip this "publicizing" list.
Done.
> >END RATIONALE.
>
> >CHARTER: comp.distributed
>
> >Although the name "comp.distributed" has been chosen due to its
> >familiarity and convenience, the group is to be broader than just
> ^^^^^^^^^
> the discussion in this
> unmoderated newsgroup
> >those topics traditionally regarded as "distributed computing".
Done.
ru.ig...@usask.ca
>future than the past, while still remembering the past :-).
I really wouldn't call this top posting since you are addressing
a different consideration than the quoted text. Besides, you
did intersperse comments within the context of the quoted text.
So if anyone complains about top posting in this message, spank
them.
[snip]
>I believe that this is, in general, a small enough modification not to
>require another RFD.
[snip]
>If we missed something, I would appreciate it if someone
>would bring it to our attention now.
I wouldn't mind seeing the updated draft of the sections in question,
just here, just to see the changes in context, especially if some of
the changes were moving text around. I'd like to believe my suggestions
make sense, but I can never count on it.
David C. DiNucci
> David C. DiNucci <da...@elepar.com> wrote:
> >I believe that this is, in general, a small enough modification not to
> >require another RFD.
> [snip]
> >If we missed something, I would appreciate it if someone
> >would bring it to our attention now.
>
> I wouldn't mind seeing the updated draft of the sections in question,
> just here, just to see the changes in context, especially if some of
> the changes were moving text around. I'd like to believe my suggestions
> make sense, but I can never count on it.
As you wish. Nearly all sections underwent minor changes, so here is
the information just as we are intending to enter it in the Proponent
Questionnaire. Readers please note: This is NOT a request for
discussion, nor a call for votes, nor anything else official, it is
simply an answer to the above request.
I am currently attempting to find out (from the other proponent and
piranha) whether I am correct in naming piranha a "Mentor" in the
questionnaire. I have not included the mailing lists we discussed
earlier because they appear all to require membership for posting, so
we'll handle pointers manually. Regarding the G5 problem, I have been
advised by Bill Aten to address these issues with the vote-taker once
one is assigned, so for now I am leaving the list of newsgroups as it
was originally.
I am hoping to submit this ASAP, so prompt feedback (if any) would be
appreciated.
Thanks,
-Dave
-----------------------------------------------------------------
David C. DiNucci Elepar Tools for portable grid,
da...@elepar.com http://www.elepar.com parallel, distributed, &
503-439-9431 Beaverton, OR 97006 peer-to-peer computing
=============================================================
SUMMARY:
The first RFD was launched under the name "comp.p2p-grid", and that
received several suggestions, primarily that the charter be broadened
and that the name be changed to "comp.distributed". The name,
rationale, and charter for the 2nd RFD (i.e. for comp.distributed) were
generally well-received. Questions centered around (a) whether it would
have sufficient traffic to justify its existence, (b) whether it would
be better off moderated (primarily to avoid spam), and (c) whether the
the group was likely to serve largely as a home base for people to
criticize other mailing lists or projects involved in this field.
Regarding (a) we provided evidence of: significant interest in the
subject around the world (as evidenced by companies, working groups,
forums, and conferences devoted to it); significant traffic (and at
least 650 and probably upwards of 1000 members) on mailing lists
associated with the topic; and traffic on non-big-8 newsgroups devoted
to related topics (e.g. alt.sci.seti, alt.gnutella,
asu.comp.distributed,
and alt.internet.p2p). Determining existing newsgroup traffic on these
specific topics was extremely difficult, due to overloaded terms like
"grid", "p2p", and "distributed" to describe the subject matter, and
the stated past wish of some to post on these subjects but their
unwillingness to post on these issues "off-topic" on other newsgroups.
Regarding (b), no group-specific reasons were given for moderation
and the proponents believe that discussion will be much better
facilitated on an unmoderated group. Regarding (c), we were unable
to discern the precise source of those concerns, but to help limit
the potential of having them resurface, we have reworded some of the
rationale comparing the proposed newsgroup to existing mailing lists.
PROPOSED NAMES:
comp.distributed
NEWSGROUP DESCRIPTIONS:
comp.distributed Distributed Resource Sharing and Exploitation.
PROPONENT:
Proponent: Rajkumar Buyya <rajk...@csse.monash.edu.au>
Proponent: David C. DiNucci <da...@elepar.com>
Mentor: piranha <pir...@gooroos.com>
RATIONALE: comp.distributed
Networks in general, and the internet specifically, have been
evolving, from star topologies of thin clients or dumb terminals
connected to central servers, to a collection of highly connected
nodes, many having significant compute, storage, and peripherals,
along with human presence. Likewise, internet tools and protocols
have evolved from being primarily a mechanism to "push" (via email)
or "pull" (via web-browser) untyped data, into supporting more
interactive, semantic, and bi-directional relationships. These
changes have prompted different communities to (re-)explore the
potential of sharing and exploiting collections of heterogeneous,
geographically distributed resources such as computers, data, people,
and scientific instruments in a secure and consistent manner, usually
lacking any central control or authority. These efforts are often
described with terms like "peer-to-peer" ("p2p") and "grids", and
can serve to virtualize enterprizes by blurring the significance of
physical location.
Different communities tend to focus on different varieties of
resources, different overall objectives and constraints, and different
degrees of permanence of the resource collectives. For example,
"grid" communities will often consider large, semi-permanent (though
dynamically constituted) collections of world-class resources that can
be accessed much as utilities, to provide unprecedented capabilities
that enable, for example, large-scale problems in science, engineer-
ing, and commerce. "p2p" communities, on the other hand, often seek
on-demand temporary relationships between everyday personal computers,
devices, and peripherals "at the edge of the network", that help to
solve every-day problems of sharing, collaboration, and computing in
more efficient, convenient, and economical ways. Similar relation-
ships have been explored over time in areas related to human collabor-
ation, distributed data bases, distributed search, parallel and
distributed computing, web services, and hierarchical content delivery
networks.
In spite of these differences, all of these communities share a large
number of challenges as a direct result of attempting to effectively
and synergistically assemble and use these collectives of hetero-
geneous distributed resources. These challenges include:
* Lack of any central authority, leading to the potential unannounced
availability or withdrawal of resources, requiring fault tolerant
applications and complicating the discovery and scheduling of
resources.
* Heterogeneous resources, requiring methods to recognize and request
unique functionality when needed, while hiding unexploitable
resource differences behind consistent interfaces.
* Heterogeneous performance in those resources, prompting the use of
simulation and performance modeling to determine which resources to
use when.
* Heterogeneous requirements from both resource owners and end users
in terms of their objectives, quality of services, and computa-
tional economy.
* Unpredictable and dynamic network topology and properties,
requiring the ability to portably deal with differing latency and
bandwidth constraints (e.g. hiding latency while minimizing
overhead) and motivating quality of service (QoS) mechanisms.
* A complex and unpredictable concurrent environment, requiring
general approaches to program development that hide these features
while leveraging existing tools, languages, and techniques wherever
possible.
* A memory hierarchy that can extend to the memory and disk throughout
the collective, prompting a reconsideration of traditional data
storage and caching approaches.
* The potential presence of untrusted resources and/or actors,
requiring decentralized approaches to privacy, authorization,
authentication, anonymity, and the determination of levels of
acceptable risk associated with different operational modes.
* Achieving return on investment for both resource users and
providers, requiring approaches for auditable accounting and re-
imbursement as well as the consideration of cost/price as a resource
selection parameter.
* Impediments to connectivity, including firewalls and oversubscribed
scarce network resources (such as dial-in modems, and IP addresses
shared through network address translation/IP masquerading).
* Cross-organizational IT involvement, requiring flexible and
politically acceptable policies, procedures, and management tools.
* Evaluating and proposing mechanisms and policies for the protection
of intellectual property in an environment explicitly designed to
facilitate instant sharing.
* Understanding and exploiting the potential value of these resource
collectives, including effective collaboration strategies,
integration of mixed resource types into problem solving
environments, novel application areas and solution approaches
enabled by this environment, and the use of automated agents.
Already, international academic and commercial forums like:
* Global Grid Forum: <http://www.gridforum.org>
* Peer to Peer Computing WG: <http://www.p2pwg.org>
* Universal Plug-n-Play Forum <http://www.upnp.org>
* New Productivity Initiative <http://www.newproductivity.org>
have evolved to create standards and protocols for inter-operability
between heterogeneous systems providing virtual services. Recently,
infrastructure projects like the NSF Distributed TeraScale Facility
have focused even more attention, and include involvement from several
companies. Many computer and/or software vendors, large and small,
have recently announced specific projects or general priorities into
p2p and/or grids, including IBM, Intel, DSTC, Sun, and Microsoft.
Some details on these and other projects can be found at:
* http://www.gridcomputing.com/
* http://www.computer.org/dsonline/gc/index.htm
* http://www.peertal.com/
* http://www.nwfusion.com/
* http://www.peerintelligence.com/
* http://www.openp2p.com/
Some existing newsgroups, like comp.parallel and comp.sys.super,
touch on specialized aspects of this topic, and will continue
to do so. However, this new group will allow more general
discussions, as well as serve as a focal point for considering the
inter-relationships, interactions, and synergies when combining
these separate technologies. In addition, over 20 existing mailing
lists relate to comp.distributed topic areas (with sign-ups currently
totalling at least 650 and likely over 1000), but they generally focus
on discussion and management of specific group projects and priorities,
and/or tend to segregate p2p and grid communities, even on topics
shared by those communities. The new newsgroup is expected to
complement these lists by providing a very broad venue for discussion
and interaction, while addressing the natural scaling problems which
would prevent these mailing lists from individually expanding their
charters to encompass all of the stated goals of this newsgroup.
CHARTER: comp.distributed
Although the name "comp.distributed" has been chosen due to its
familiarity and convenience, the discussion in this unmoderated
newsgroup is to be broader than just those topics traditionally
regarded as "distributed computing". Specifically, topics are to
include any unique issues relating to the creation and exploitation
of collectives of geographically distributed and potentially
heterogeneous resources such as computers, data/information sources,
peripherals, instruments, and humans. Appropriate areas of discussion
in this context would include (but are not limited to):
* discovering, scheduling/brokering, and accessing remote resources
* exploitation of heterogeneous resources
* resource management, scheduling, and computational economy
* portable/adaptable communication substrates
* quality of service approaches
* portable program development tools, languages, techniques
* data management tools and techniques
* exploitation of distributed memory hierarchy
* decentralized security
* practical accounting, reimbursement, and business & revenue models
* overcoming impediments to wide-area connectivity
* cross-organizational policy issues and ways to address them
* mechanisms and policies for intellectual property
* programming tools, environments, and languages
* applications, collaboration, and distributed agents
* simulation and performance modelling
* comparisons of grid and p2p, and issues unique to each
* events, surveys, news and general announcements
It is expected that additional 3rd-level subgroups addressing some of
these topics or others may be created as dictated by the volume and
cohesiveness of resulting message traffic.
MODERATOR INFO: comp.distributed
GATEWAYS:
NEWSGROUPS:
Newsgroups: news.announce.newgroups,news.groups,
comp.arch,
comp.parallel,
comp.parallel.pvm,
comp.sys.super,
comp.client-server
NEW CROSSPOSTS:
none
MODERATED NEWSGROUPS:
comp.parallel
POINTER MESSAGES:
alt.internet.p2p
asu.comp.distributed
alt.gnutella
alt.sci.seti
MAILING LISTS:
ru.ig...@usask.ca
>Some existing newsgroups, like comp.parallel and comp.sys.super,
>touch on specialized aspects of this topic, and will continue
>to do so. However, this new group will allow more general
>discussions, as well as serve as a focal point for considering the
>inter-relationships, interactions, and synergies when combining
>these separate technologies. In addition, over 20 existing mailing
>lists relate to comp.distributed topic areas (with sign-ups currently
>totalling at least 650 and likely over 1000), but they generally focus
>on discussion and management of specific group projects and priorities,
>and/or tend to segregate p2p and grid communities, even on topics
>shared by those communities. The new newsgroup is expected to
>complement these lists by providing a very broad venue for discussion
>and interaction, while addressing the natural scaling problems which
>would prevent these mailing lists from individually expanding their
>charters to encompass all of the stated goals of this newsgroup.
Looks fine, except when compared to the 2nd RFD:
>>Although over 20 discussion mailing lists operated by individuals or
>>institutions exist, they are generally intended for discussion of
>>specific group priorities, and strongly segregate p2p and grid
>>communities, even when addressing similar issues. Another concern is
>>that mailing lists are likely to generate large volume of email for
>>members; therefore, moderators will often discourage use of these
>>lists for general or controversial discussion, and many prospective
>>participants feel discouraged from subscribing, do not become members,
>>and do not join important topical discussions. We believe that having
>>a newsgroup where people can participate in discussions of their own
>>choosing, when they want, without getting swamped with emails, will
>>help overcome these limitations and will encourage discussion and
>>dissemination without the need of explicit membership. While some
>>existing newsgroups, like comp.parallel and comp.sys.super, touch on
>>some specialized aspects of this topic, and will continue to do so,
>>this new group will serve as a focal point for considering the inter-
>>relationships, interactions, and synergies when combining these
>>separate technologies.
Was that the last RFD version? Is that enough of a difference to
require a new RFD? It's a change for the better, but it does
read differently enough and has a lot chopped out from the old
version. I think a new RFD is required, at least, so that it
can't be argued that folks voted on their impression of the
(different) last proposal during discussion (rather than actually
reading the CFV text), but moreso that folks get a chance to
comment in the context of the degree of change in that section
(even if it is to say it's fine).
>CHARTER: comp.distributed
I also missed some tense problems. Present tense prefered here.
>Although the name "comp.distributed" has been chosen due to its
>familiarity and convenience, the discussion in this unmoderated
>newsgroup is to be broader than just those topics traditionally
^^^^^
delete
>regarded as "distributed computing". Specifically, topics are to
^^^^^^
delete
>include any unique issues relating to the creation and exploitation
>of collectives of geographically distributed and potentially
>heterogeneous resources such as computers, data/information sources,
>peripherals, instruments, and humans. Appropriate areas of discussion
>in this context would include (but are not limited to):
^^^^^
delete
>It is expected that additional 3rd-level subgroups addressing some of
>these topics or others may be created as dictated by the volume and
>cohesiveness of resulting message traffic.
I'm having second thoughts about including this paragraph.
It's pretty much a given.
I wouldn't consider any of these suggested changes in the charter
to require a new RFD.
David C. DiNucci
>
> David C. DiNucci <da...@elepar.com> wrote:
>
> >Some existing newsgroups, like comp.parallel and comp.sys.super,
>
> Looks fine, except when compared to the 2nd RFD:
>
> >>Although over 20 discussion mailing lists operated by individuals or
> >> ....
>
> Was that the last RFD version? Is that enough of a difference to
> require a new RFD? It's a change for the better, but it does
> read differently enough and has a lot chopped out from the old
> version. I think a new RFD is required, at least, so that it
> can't be argued that folks voted on their impression of the
> (different) last proposal during discussion (rather than actually
> reading the CFV text), but moreso that folks get a chance to
> comment in the context of the degree of change in that section
> (even if it is to say it's fine).
The new language contains some wording changes as outlined by you
word-for-word, other changes removing open judgements like "large
volume" that you objected to, and other changes that simply weaken any
language that might somehow be construed as critical of mailing lists,
as you alone seemed to imply it did. How on Earth could that require a
new RFD? Who could you imagine would disagree now that wouldn't have
disagreed before? You?
On Oct 29, I specifically asked if the changes requested so far would
require a new RFD, because we wanted to go to CFV. I got a "no they
don't" response from Jay Denebeim, and you responded to the post without
commenting on the question. Now, a week later, you say they do, because
of the precise wording I have used to incorporate some of YOUR
recommendations.
Rather than have a new RFD, please just suggest how you imagined that I
was planning to reword it when you believed that another RFD was not
required, and I'll incorporate that wording instead of my new wording.
Then we'll all be happy, with no new RFD.
...
> I wouldn't consider any of these suggested changes in the charter
> to require a new RFD.
I don't understand what difference it makes whether "these" changes or
"those" changes require an RFD, when you state above "I think a new RFD
is required".
We are 27 days past the first RFD, 12 days past the last, the only
changes made are those to specifically address your concerns, I received
an answer a week ago (with no dissension) that further RFDs would not be
required, and several days have gone by without any substantive concerns
at all. After all of this, if you demand a new RFD, the next go around
will be missing at least one of its proponents.
-- Dave
David C. DiNucci
> Rather than have a new RFD, please just suggest how you imagined that I
> was planning to reword it when you believed that another RFD was not
> required, and I'll incorporate that wording instead of my new wording.
> Then we'll all be happy, with no new RFD.
Hello? Should we now be delaying the comp.distributed CFV further
because someone makes a "drive by" demand for a new RFD and then
disappears when asked to justify and/or correct their claims?
I do not know what authority, if any, this "ru" person has to tie up the
RFD/CFV process. I understand that this process is run largely by
volunteers, but I am also volunteering my time on this, and have made a
good faith effort to allow people to voice realistic concerns. If
people cannot explain their concerns, I have no choice but to consider
those concerns unworthy of any further delays, so I intend to submit the
CFV questionnaire information as it was posted here recently, with the
few explicit wording changes (correcting tense) just recommended by
"ru", and one possible additional change: One proponent has requested
that the last word of the group description
comp.distributed Distributed Resource Sharing and Exploitation.
...be changed to "Applications" i.e.
comp.distributed Distributed Resource Sharing and Applications.
This does not change the meaning (i.e. applying distributed resources is
the same as exploiting them, but some find the new language more
natural).
If anyone objects to these few modifications, or feels that they would
require a new RFD, we will back out and leave the questionnaire
responses as they were recently posted.
Thanks,
Jay Denebeim
David C. DiNucci <da...@elepar.com> wrote:
>I do not know what authority, if any, this "ru" person has to tie up the
>RFD/CFV process.
Nobody has any authority. Ru is just posting what he thinks as do the
rest of us. He usually is right on the money though, although I don't
know what you're referring to here. So, I'd give what he says the
consideration you'd give someone who pretty much knows what he's
talking about and then decide for yourself.
You get to make the call when to start the CFV process by filling out
the PQ. You can do as many or as few RFDs as you wish up to that
point. About the only thing enforced is if you change something
substantive between your last RFD and what you turn in for the CFV the
news.announce.newgroups moderator may insist you do another RFD.
Other than that it's totally up to you.
Heck, we've even seen some people post an RFD, ignore all discussion
and then ask for a CFV the minimum 21 days later. I don't believe any
of them have passed though.
Graham Drabble
news:9s9olh$896$2...@dent.deepthot.org:
> Heck, we've even seen some people post an RFD, ignore all
> discussion and then ask for a CFV the minimum 21 days later. I
> don't believe any of them have passed though.
There was rec.games.planeterion that had no discussion at all until
after the PQ was submitted. Somehow everyone seemed to miss it and the
proponent didn't say anything. It passed but I think it got a lot of
nos.
--
Graham Drabble
If you're interested in what goes on in other groups or want to find
an interesting group to read then check news.groups.reviews for what
others have to say or contribute a review for others to read.
Brian Mailman
> ...You can do as many ... RFDs as you wish up to that point [of filling out the PQ].
Ummm.... I think the anti-fascism groups may have changed that.
B/
Shrisha Rao
>>
>> ru.ig...@usask.ca wrote:
>> >...
>> Rather than have a new RFD, please just suggest how you imagined that I
>> was planning to reword it when you believed that another RFD was not
>> required, and I'll incorporate that wording instead of my new wording.
>> Then we'll all be happy, with no new RFD.
>
>Hello? Should we now be delaying the comp.distributed CFV further
>because someone makes a "drive by" demand for a new RFD and then
>disappears when asked to justify and/or correct their claims?
Almost everything you read here is free, non-binding advice. You
don't have to delay anything. A few people may vote NO because
you didn't do what they thought was right, but I'd just ignore that.
If your proposal is at risk from these few, then it probably won't
pass anyway. In rare circumstances the UVV and/or Tale may step in
and object to the way you're doing something, but I don't think that
likely in your case (and you have the word of someone who's tested the
boundaries more often than anyone else).
I personally don't care that much about the precise wordings of
charters or such for unmoderated groups because they don't matter in
practice. Many news.groupies however don't (want to) realize this,
and/or take themselves too seriously.
Good luck, and don't lose sleep (or your temper) over this.
Regards,
Shrisha Rao
>-- Dave
--
David C. DiNucci
>
> In article <3BE84D49...@elepar.com>,
> David C. DiNucci <da...@elepar.com> wrote:
>
> >I do not know what authority, if any, this "ru" person has to tie up the
> >RFD/CFV process.
>
> Nobody has any authority. Ru is just posting what he thinks as do the
> rest of us. He usually is right on the money though, although I don't
> know what you're referring to here.
I think I was pretty clear in the message you're responding to. Not
only is Ru saying that a new RFD is required based on the most minor of
changes (which were, in fact, all requested by Ru), we requested to be
told a week ago if these might justify the need for another RFD. You
said "no", and Ru said nothing. There are several conferences around
the world on these topics in the next few weeks/months, and it would be
good to have the group available for that traffic ASAP.
> You get to make the call when to start the CFV process by filling out
> the PQ. You can do as many or as few RFDs as you wish up to that
> point. About the only thing enforced is if you change something
> substantive between your last RFD and what you turn in for the CFV the
> news.announce.newgroups moderator may insist you do another RFD.
In case that moderator (who, then, seems to have some authority) should
refer to this group to determine whether we did "change something
substantive", here is a point-by-point illustration that we did not.
[Quoting in the next several paragraphs refers to the new rationale (>)
and the original rationale (> >).]
First two sentences of the new paragraph:
>Some existing newsgroups, like comp.parallel and comp.sys.super,
>touch on specialized aspects of this topic, and will continue
>to do so. However, this new group will allow more general
>discussions, as well as serve as a focal point for considering the
>inter-relationships, interactions, and synergies when combining
>these separate technologies.
were extracted directly from the last line of the original paragraph
with minor wording changes as requested specifically by Ru, and moved to
the top, as requested specifically by Ru. The original text read:
> > ...While some
> >existing newsgroups, like comp.parallel and comp.sys.super, touch on
> >some specialized aspects of this topic, and will continue to do so,
> >this new group will serve as a focal point for considering the inter-
> >relationships, interactions, and synergies when combining these
> >separate technologies.
The third sentence in the new paragraph:
>In addition, over 20 existing mailing
>lists relate to comp.distributed topic areas (with sign-ups currently
>totalling at least 650 and likely over 1000), but they generally focus
>on discussion and management of specific group projects and priorities,
>and/or tend to segregate p2p and grid communities, even on topics
>shared by those communities.
is a rephrasing of the first sentence of the original paragraph (below),
with additional information on number of subscribers as requested by Ru,
which were cited on this group, and easily verifiable (e.g. the yahoo
decentralization mailing list reports 652 subscribers):
> >Although over 20 discussion mailing lists operated by individuals or
> >institutions exist, they are generally intended for discussion of
> >specific group priorities, and strongly segregate p2p and grid
> >communities, even when addressing similar issues.
The fourth and final sentence of the new paragraph:
> ... The new newsgroup is expected to
>complement these lists by providing a very broad venue for discussion
>and interaction, while addressing the natural scaling problems which
>would prevent these mailing lists from individually expanding their
>charters to encompass all of the stated goals of this newsgroup.
is a rephrasing and softening of remainder of the original paragraph
(below), to address Ru's concerns that it somehow suggested that the
newsgroup would be critical of mailing lists, and his statement that
"Really, the excuse given [...] is always true for any mailing list,
it's just a question of degree.":
> > ... Another concern is
> >that mailing lists are likely to generate large volume of email for
> >members; therefore, moderators will often discourage use of these
> >lists for general or controversial discussion, and many prospective
> >participants feel discouraged from subscribing, do not become members,
> >and do not join important topical discussions. We believe that having
> >a newsgroup where people can participate in discussions of their own
> >choosing, when they want, without getting swamped with emails, will
> >help overcome these limitations and will encourage discussion and
> >dissemination without the need of explicit membership.
So, the only negative feedback that a new RFD could invite that the old
one didn't is that there is now too much information in the rationale,
or the rationale isn't critical enough of mailing lists, neither of
which has merit.
Jay Denebeim wrote:
> Heck, we've even seen some people post an RFD, ignore all discussion
> and then ask for a CFV the minimum 21 days later. I don't believe any
> of them have passed though.
I think you'll agree that that bears no resemblance to this instance,
where we have gone out of our way to solicit and respond to input, we
have posted (and in many cases, reposted, to get around the G5 problem)
the RFD to many newsgroups and mailing lists, and have already gone a
week past the minimum 21 days to ensure the opportunity to comment.
Lacking further dely, the next thing you are likely to see on this
subject is a CFV.
piranha
> because someone makes a "drive by" demand for a new RFD and then
> disappears when asked to justify and/or correct their claims?
no.
> I do not know what authority, if any, this "ru" person has to tie up the
> RFD/CFV process.
none. nobody is tying this process up. ru gives advice. you can
take it or leave it. the worst that can happen is that you'd
garner his "no" vote but i haven't seen him say anything that
would lead me to think minor wording changes would incline him to
do so.
that is true for everyone here. the only person who can really
stop a vote would be tale, and you have no reason whatsoever to
worry about that.
you also have no reason to worry about massive no votes due to
serious technical reasons.
so the decision when to go to CFV is up to you, since the minimum
discussion time period has passed.
> If anyone objects to these few modifications, or feels that they would
> require a new RFD, we will back out and leave the questionnaire
> responses as they were recently posted.
no, they do not require a new RFD.
--
-piranha
Jay Denebeim
>Jay Denebeim wrote:
>> Heck, we've even seen some people post an RFD, ignore all discussion
>> and then ask for a CFV the minimum 21 days later. I don't believe any
>> of them have passed though.
>I think you'll agree that that bears no resemblance to this instance,
Definately. You've been fine as a proponant.
David C. DiNucci
> Definately. You've been fine as a proponant.
>
> Jay
Thanks all for the input. (Yes, Ru too. As illustrated here earlier,
most every wording clarification Ru suggested was incorporated.) If I
sense a problem with this process, it is not with an unwillingness or
inability of the participants to provide useful feedback. It is with
the inability to define terms like "significant changes". Since they
can't be defined formally and objectively, it means that someone
somewhere gets to define them informally and subjectively--but who that
person is (those people are), or when the decision can be made, is not
clear at all.
However, maybe the frustration of not knowing is less than the
frustration of dealing with the beauracracy that would be required to
improve it.
Oh well. The CFV machinery has hopefully been set in motion.
Jay Denebeim
David C. DiNucci <da...@elepar.com> wrote:
>It is with the inability to define terms like "significant changes".
Oh, they have to be *really* significant changes, like changing the
name of the group.
David C. DiNucci
>
> In article <3BE97A44...@elepar.com>,
> David C. DiNucci <da...@elepar.com> wrote:
>
> >It is with the inability to define terms like "significant changes".
>
> Oh, they have to be *really* significant changes, like changing the
> name of the group.
That may be, but it is not at all apparent from the documentation. For
example, here is some text associated with the Proponent Questionnaire:
> Proponents are expected to react to issues raised during the RFD
> process and incorporate changes based upon this process; if there
> are substantial changes from the last RFD, then you need to
> submit another RFD, not this questionnaire. Review Question #1
> (in the "before you start" section below) to ensure that you're
> in compliance with this requirement.
Question 1 then contains this text:
> NOTE: If, based on the input you've received, it appears that
> changes to the name(s) (including typos), charter(s), or
> moderator information of any of the proposed newsgroup(s) are
> necessary, you must make those changes and issue another RFD and
> allow discussion of the new RFD before submitting this form.
I read this as saying that one MUST issue a new RFD if changes are made
to the charter. Any changes? How about the tense changes that Ru
suggested? I assumed they were OK. How about the deletion of the last
paragraph of the charter that Ru suggested? I decided not to chance it,
if it really was just satisfying a personal preference for simplicity.
How about changing the last word of the newsgroup description, as we
recommended (from "Exploitation" to "Applications")? Note that the
instructions say "including typos" in one spot, suggesting that even
very small changes in some places can require a new RFD. I ended up
using my own judgement, which I hope will match the moderator's. And if
it doesn't, I'll hope to handle it through discussion.
Rajkumar Buyya
ru.ig...@usask.ca wrote:
....
...
> >include any unique issues relating to the creation and exploitation
> >of collectives of geographically distributed and potentially
> >heterogeneous resources such as computers, data/information sources,
> >peripherals, instruments, and humans. Appropriate areas of discussion
> >in this context would include (but are not limited to):
> ^^^^^
> delete
>
> >It is expected that additional 3rd-level subgroups addressing some of
> >these topics or others may be created as dictated by the volume and
> >cohesiveness of resulting message traffic.
>
> I'm having second thoughts about including this paragraph.
> It's pretty much a given.
>
> I wouldn't consider any of these suggested changes in the charter
> to require a new RFD.
In that case we go for CFV!
Thanks
Raj