comp....andrew...
Eliot Lear
>
> No agreement on the name. There's no 'comp.sw-sys' to stick the group on.
> I called for a vote on a comp.sw.misc to provide such a place, but after
> receiving mail from one of the architects of the great renaming that the
> comp.sys hierarchy was intended to incluide software systems I'm not sure if
> I should start the vote on comp.sw.misc or just create comp.sys.andrew?
>
NO. There exists an inet group called comp.soft-sys.andrew. The
group can easily be extended to accomodate the rest of the net. Also,
one of the engineers of the great renaming told me just the opposite.
Let's take a look and see who is more apt to be correct:
comp.sys.amiga, comp.sys.amiga.tech, comp.sys.apollo, comp.sys.apple,
comp.sys.atari.8bit, comp.sys.atari.st, comp.sys.att, comp.sys.cbm,
comp.sys.cdc, comp.sys.celerity, comp.sys.dec, comp.sys.dec.micro,
comp.sys.encore, comp.sys.handhelds, comp.sys.hp, comp.sys.ibm.pc,
comp.sys.ibm.pc.digest, comp.sys.ibm.pc.rt, comp.sys.intel,
comp.sys.intel.ipsc310, comp.sys.m6809, comp.sys.m68k, comp.sys.m68k.pc,
comp.sys.mac, comp.sys.mac.digest, comp.sys.mac.hypercard,
comp.sys.mac.programmer, comp.sys.masscomp, comp.sys.misc,
comp.sys.northstar, comp.sys.nsc.32k, comp.sys.proteon, comp.sys.pyramid,
comp.sys.ridge, comp.sys.sequent, comp.sys.sgi, comp.sys.sun, comp.sys.super,
comp.sys.tahoe comp.sys.tandy, comp.sys.ti, comp.sys.ti.explorer,
comp.sys.transputer, comp.sys.workstations, comp.sys.xerox, comp.sys.zenith,
comp.sys.zenith.z100, comp.sys.next
Hmmm.. I count a grand total of 1 software system in there.
Nope. Let's extend comp.soft-sys. comp.sw doesn't convey any
additional information. Just about everything we talk about deals
with software. I've heard an argument for one better, and that would
be comp.software-systems, which would make it easier for non english
speaking people to understand. Anyway, when we finally do agree on a
name (probably just in time to catch the creation of EEC), I suggest
we move comp.emacs and into that more fitting home.
In the meantime, unless there are some real (*) objections, maybe spaf
would be so kind as to issue a ``world'' newgroup for
comp.soft-sys.andrew. As has been stated, there was a successful
vote, and we can always change the name again, later, should there
really be that much interest.
(*) A real objection would be a good reason to have people go to the
trouble of renaming the existing group.
--
Eliot Lear
[le...@net.bio.net]
Eliot Lear
It makes absolutely no sense at all to create any part of hierarchy
that doesn't convey additional information. Maybe we should call into
question the naming of comp.sw.components. Even if you still had it
in your head to create that groups, I would argue strenuously that
``software'' should NOT be shortened to sw. Why make the name any
more meaningless than it already is? Even soft-sys fails in this
respect, but it exists already.
Since I don't recall the charter for comp.sw.components, I can't argue
for or against.
As far as how comp.soft-sys.andrew was created, Erik Fair is the one
person who is within his rights to create groups without votes, but
the distribution of those groups is to be restricted to ``inet''. His
right to do so is not open to debate in this forum. Furthermore,
news.groups could not agree on a name for the andrew group. What's
more, Nathaniel's votes were continually called into question.
Therefore, your objection is invalid.
--
Eliot Lear
[le...@net.bio.net]
Peter da Silva
> NO. There exists an inet group called comp.soft-sys.andrew. The
> group can easily be extended to accomodate the rest of the net. Also,
> one of the engineers of the great renaming told me just the opposite.
Erik Fair? The guy who created comp.soft-sys.andrew by imperial fiat
despite it not being the name voted on?
> Nope. Let's extend comp.soft-sys. comp.sw doesn't convey any
> additional information.
Maybe, maybe not. It's more general (for example, comp.sw.components makes
sense, but comp.soft-sys.components doesn't), and Bill Wolfe is preparing
to create it... with a bit more (just a bit more) justification: namely
that he has a real group voted on in that hierarchy.
> (*) A real objection would be a good reason to have people go to the
> trouble of renaming the existing group.
How about "the existing group was not voted on, and the group voted on (either
comp.andrew or comp.sys.andrew) was never created"?
--
Peter da Silva, Xenix Support, Ferranti International Controls Corporation.
Business: uunet.uu.net!ficc!peter, pe...@ficc.uu.net, +1 713 274 5180.
Personal: ...!texbell!sugar!peter, pe...@sugar.hackercorp.com.
Peter da Silva
> It makes absolutely no sense at all to create any part of hierarchy
> that doesn't convey additional information. Maybe we should call into
> question the naming of comp.sw.components.
Maybe we should. I attempted to do so at the time of the call to votes.
You were conspicuously silent at that time, as I recall.
> Even if you still had it
> in your head to create that groups, I would argue strenuously that
> ``software'' should NOT be shortened to sw. Why make the name any
> more meaningless than it already is? Even soft-sys fails in this
> respect, but it exists already.
It's not my choice either. Take it up with Bill Wolfe.
> As far as how comp.soft-sys.andrew was created, Erik Fair is the one
> person who is within his rights to create groups without votes, but
> the distribution of those groups is to be restricted to ``inet''. His
> right to do so is not open to debate in this forum.
I never said otherwise. But, as you say, "the distribution of those groups is
to be restricted to ``inet''". And his right to create a group does not in any
way invalidate anyone's right to point out flaws in that group's name, charter,
or anything else.
> Furthermore,
> news.groups could not agree on a name for the andrew group. What's
> more, Nathaniel's votes were continually called into question.
I see. a flawed vote (that passed even after the objectionable votes were
removed from the count) is less valid than no vote at all. I believe that
the group called for, voted on, and passed was actually "comp.andrew".
> Therefore, your objection is invalid.
You have yet to show that. INET != USENET. There is no-one entitled to create
groups on usenet without a vote.
Eliot Lear
> You were conspicuously silent at that time, as I recall.
Sorry, I have a life. I have my own distribution to maintain
(bionet), which is actually part of a real job.
As far as proving that INET != USENET, as I said in personal mail,
there need not be such a proof. What there needs to be is a reason
to break away from what exists in a rather prevalent hierarchy.
Please give me a reason why we should cause the confusion and the
inconvenience necessary to change the existing group.
--
Eliot Lear
[le...@net.bio.net]
Eliot Lear
comp.sw. I believe it to be a vast mistake, as the added two letters
(sw) convey no additional meaning, and are not particularly useful for
classifying information.
I respectfully request that no group in the comp.sw tree be created
until we can come to agreement on a subhierarchy (that I *do* believe
we need). Any hierarchy chosen should aid the end user in
categorizing both data to be found and data to be posted.
Finally, we should consider adoption of the soft-sys hierarchy, given
that it already exists in the inet distribution.
--
Eliot Lear
[le...@net.bio.net]
Rich Goldschmidt
vote has not been created. My understanding of group creation guidlines was
that the discussion period was intended to resolve issues like what the name
would be, and where in the hierarchy it would go. If people don't like it they
vote no. Having voted yes, the group should be created. If people later
vote to create some new segment of the hierarchy, they can also vote to
rename an existing group. But to ignore a positive vote because someone
suggests after the fact that it should have a different name seems to violate
both the letter and the spirit of all the posted guidelines I have seen.
Please create comp.andrew or comp.windows.andrew or comp.soft-sys.andrew
or any other group that will accomodate these discussions! It has already
been approved by a vote...
Rich Goldschmidt
uunet!rlgvax!golds
William Thomas Wolfe,2847,
> Oh, so I do call into question the creation of a hierarchy known as
> comp.sw. I believe it to be a vast mistake, as the added two letters
> (sw) convey no additional meaning, and are not particularly useful for
> classifying information.
Not true; comp means computer, sw means software, computer /= software,
hence additional meaning, and additional resolution of classification.
Bill Wolfe, wtw...@hubcap.clemson.edu
Eliot Lear
I've been waiting for someone to say exactly what you said:
> Not true; comp means computer, sw means software, computer /= software,
> hence additional meaning, and additional resolution of classification.
Oh come on. Out of the existing hierarchy, we could classify the
following as software:
compilers, text, sources, windows, unix, emacs, editors, binaries, os,
databases, lang, mail, bugs, fonts, (1/2) graphics, (1/2) parallel,
(1/2) protocols.
And then there are other groups where software is often discussed.
How about comp.sys.mac.hypercard?
Even if you found some meaning in a software hierarchy, I would still
object to you shortening software to sw. Why bother? Again, all you
do is make the name more meaningless.
--
Eliot Lear
[le...@net.bio.net]
William Thomas Wolfe,2847,
>> Not true; comp means computer, sw means software, computer /= software,
>> hence additional meaning, and additional resolution of classification.
>
>
> compilers, text, sources, windows, unix, emacs, editors, binaries, os,
> databases, lang, mail, bugs, fonts, (1/2) graphics, (1/2) parallel,
> (1/2) protocols.
>
> And then there are other groups where software is often discussed.
> How about comp.sys.mac.hypercard?
The purpose of the software hierarchy is not to discuss particular
software systems or applications, but software in general;
software engineering, software components, things like that.
> Even if you found some meaning in a software hierarchy, I would still
> object to you shortening software to sw. Why bother? Again, all you
> do is make the name more meaningless.
If you oppose abbreviated names in hierarchies, there is certainly
plenty to oppose, comp being an excellent place to start. However,
until the next Great Renaming (at which point you will have an
opportunity to advance such an argument, and I may well support it),
I think it appropriate to be consistent with ai, os, sf, edu, etc.
Bill Wolfe, wtw...@hubcap.clemson.edu
Peter da Silva
> Please give me a reason why we should cause the confusion and the
> inconvenience necessary to change the existing group.
Because a new level to the usenet hierarchy is too important for hasty
and short-sighted action. I agree, comp.sw is ugly. Comp.software would be
much more logical. The problem is that comp.soft-sys is not only ugly,
but it's overly restrictive.
comp.soft-sys.andrew OK. Ugly, but makes sense.
comp.soft-sys.components No sense here.
comp.soft-sys.engineering Nope.
And would you move other software systems groups from comp.sys and comp
into the new hierarchy? If not, why not comp.andrew or comp.sys.andrew?
If so, why is the hassle of moving comp.soft-sys.andrew so much greater?
Maybe it's time for another Great Renaming?
And if not, comp.sys is fine.
Would it be against netiquette to quote personal correspondence here?
Eliot Lear
I suppose we want to discuss what we want to carry in said new
hierarchy. As Dave Lawrence said and as I have inferred, software is
a big word filled with vagueness. One minor problem that I noticed
when I read Dave's message was that I neglected to mention one of my
goals, which was to have a place to put big packages like Andrew and
EMACS (it's not just an editor, it's an adventure ;-). Yet, Bill, you
want to have a place to discuss the concept of software. For a
hierarchy to fit both of our types of groups, I claim that its scope
would have to be unacceptably large and require our little
monstrosity.
Given that, we need two hierarchies. The question: can we put
comp.sw.components in an existing hierarchy? My immediate thought is
comp.theory.software-components. Or we could go for
comp.theory.software.{components,engineering,...}. If that occurs,
then comp.software-systems or comp.soft-sys or some such thing would
make sense.
WhatdyaThink?
--
Eliot Lear
[le...@net.bio.net]
David C Lawrence
Eliot> Even if you found some meaning in a software hierarchy, I would still
Eliot> object to you shortening software to sw. Why bother? Again, all you
Eliot> do is make the name more meaningless.
In <53...@hubcap.clemson.edu> billwolf%hazel.cs.c...@hubcap.clemson.edu
(William Thomas Wolfe,2847,) writes:
Bill> If you oppose abbreviated names in hierarchies, there is certainly
Bill> plenty to oppose, comp being an excellent place to start. However,
Bill> until the next Great Renaming (at which point you will have an
Bill> opportunity to advance such an argument, and I may well support it),
Bill> I think it appropriate to be consistent with ai, os, sf, edu, etc.
I am of Eliot's opinion on this for two reasons:
a) "sw" is not an inherently obvious abbreviation for "software", even
for people that spend most of their time around computers. Even
though Bill says that sw and hw have been around for ages as
abbreviations (which I do not attempt to dispute) they are -nothing-
like AI, OS and SF, which people say in regular conversation by merely
pronouncing the letters. "edu" is recognized as educat* in even
non-computer circles as are "alt", "sci", "rec", "misc", et al. "sw is
not nearly as obvious, as a quick straw poll locally just proved.
b) The complaint I have seen about comp.soft-sys.andrew is not that
soft-sys is too long, but that comp.soft-sys.components makes no
sense. Right, comp.software.components does make more sense. Too
long? "software" is as long as "soft-sys" and most modern news
readers either have auto-completion (ie, GNUS) or can go to a group
based on a smaller portion of the name (ie, "o" in rn). And anytime
people get that lazy about typing a name it usually just ends up as
z.y.z -- r.h.f is a good example.
My problem with comp.software is that it loses that inherent link with
entire software systems and just ends up looking like an extremely
broad category for anything that is software. This is not really a
problem though as the well-informed and educated readers of news.groups
would be sure not to put a group there which really didn't fit the
software-system idea. (1/2 :-)
Dave
--
ta...@rpitsmts.bitnet, tale%m...@itsgw.rpi.edu, ta...@pawl.rpi.edu
William Thomas Wolfe,2847,
> [Eliot sent me a copy of his recent article via e-mail, which I
> responded to not knowing that it was also posted, so I'm also posting]
I think there is a comp.sys hierarchy which was designed to handle
large software systems such as andrew and emacs; probably the names
comp.sys.andrew and comp.sys.emacs would then be appropriate for the
respective groups. Greg Woods posted something a week or two back
indicating that comp.sys was intended to cover hw, hw/sw, and sw
systems, and at that point I thought the whole andrew question had
been settled. Components doesn't fit because components are atomic
(not a system, at least externally), and also components are being
discussed more or less as an abstract proposition, although some
discussion of specific components will inevitably arise from time
to time. I really think there is a need for a place to discuss
software as a general proposition, and that the sw hierarchy will
provide a happy home for the groups that want to do that.
Perhaps the reason comp.sys isn't catching on too well is that
something like comp.env (env for environment) may have been more
along the lines of what you had in mind?
Bill Wolfe, wtw...@hubcap.clemson.edu
Eliot Lear
Well I disagree with Greg Woods regarding the meaning of comp.sys.
One might argue that we blew it in the great renaming, and that there
should have been a comp.hardware and comp.software, but I am rather
glad we didn't take that route, because I doubt the usefulness of such
names.
Regardless of what the reason WAS, there exists a good reason to keep
everything but hardware OUT of that group. People EXPECT to find
hardware computer systems in comp.sys. That hierarchy has partitioned
the comp tree quite nicely in a rather intuitive manner. It is just a
little unfortunate that it has another little relatively meaningless
and vague name.
env suffers from two flaws - it is unnecessarily short (we should
spell out environment, if anything) and the word environment is still
quite vague. Remember, we are out to make the most of the newsgroup
name that we can. It is the one mandatory (topic classified) keyword
in a message.
In fact, I wouldn't object to going with comp.software-systems. The
only reason that I wouldn't want comp.software.systems is that the
word ``software'' does not stand on its own. It's just too vague to
be useful (and yes, I predict lots of abuse of such a hierarchy).
Finally, let me just emphasize what I said in my previous message:
I believe we need two different hierarchies.
--
Eliot Lear
[le...@net.bio.net]
David Wright
That is what should be created.
I still think Erik was mistaken (entitled, but mistaken) to invent the
new and awkward "comp.soft-sys.andrew" for the inet group - a name that
had not been discussed in news.groups at all. I was very unhappy with
that at the time, and said so - and locally I call the group comp.sys.andrew
in accordance with the vote (an entry in the aliases files ensures our
readers get the articles sent to comp.soft-sys.andrew).
The idea that "systems" = "hardware" should have been recognised as a
fallacy many many years ago (indeed I remember first having that argument
about 1969!). "Andrew" is a system, just as "isis" is - we already have
comp.sys.isis because that is the best place to put it: let us do the same
with Andrew.
The only sensible alternative name that I have seen is comp.env.andrew,
but that means creating a new second-level branch, and risks arguments
about comp.env.emacs too :-)
During the pre-vote debate, someone suggested "comp.windows.andrew", but
the reponse was to the effect that Andrew is a complete system with filer,
editors, mail, comms etc. etc. as well as windows. So as it is a system,
let it be called one - comp.sys.andrew.
Regards, David Wright STL, London Road, Harlow, Essex CM17 9NA, UK
d...@stl.stc.co.uk <or> ...uunet!mcvax!ukc!stl!dww <or> PSI%234237100122::DWW
Living in a country without a written constitution means having to trust in
the Good Will of the Government and the Generosity of Civil Servants.
Eliot Lear
Again, I rise to argue to maintain the separation between hardware and
software as there exists a natural partition that the average
individual will use. I will concede that "sys" != "hardware" in real
life, and for purposes of argument, I will even grant that the
backbone was conscious of this fact when The Great Renaming took place.
I still claim that neither of those facts are important, because a
partition has formed, and we should take piety on the poor news
subscriber and not lightly ignore (destroy) it.
--
Eliot Lear
[le...@net.bio.net]
Greg Woods
> Greg Woods posted something a week or two back indicating that
> comp.sys was intended to cover hw, hw/sw, and sw systems,
I think this has been misattributed. I posted no such thing. If you still
think I did, I'd like a Message-ID to refer back to. Perhaps it's just that
something I did say has been misinterpreted.
--Greg
Peter da Silva
> Greg Woods posted something a week or two back indicating that
> comp.sys was intended to cover hw, hw/sw, and sw systems,
In article <31...@ncar.ucar.edu>, wo...@ncar.ucar.edu (Greg Woods) replies:
> I think this has been misattributed. I posted no such thing. If you still
> think I did, I'd like a Message-ID to refer back to. Perhaps it's just that
> something I did say has been misinterpreted.
I posted that message. I was told by Rick Adams (ri...@uunet.uu.net) that
he was the guy who thought up the name and that comp.sys was intended to
include software and hardware systems.
William Thomas Wolfe,2847,
>billwolf%hazel.cs.c...@hubcap.clemson.edu writes:
>> Greg Woods posted something a week or two back indicating that
>> comp.sys was intended to cover hw, hw/sw, and sw systems,
>
> I think this has been misattributed. I posted no such thing. If you still
> think I did, I'd like a Message-ID to refer back to. Perhaps it's just that
> something I did say has been misinterpreted.
The article's long gone at my site, but I do remember somebody saying
that the creators of comp.sys intended it to cover both hardware
and software systems, in almost exactly those words.
Bill Wolfe, wtw...@hubcap.clemson.edu
Eliot Lear
> The article's long gone at my site, but I do remember somebody saying
> that the creators of comp.sys intended it to cover both hardware
> and software systems, in almost exactly those words.
Need we take a survey of those who did and those who didn't? One
person from the backbone said that. It doesn't make it gospel. At
least one other said the exact opposite. Again, it is sort of useless
to argue what the sys hierarchy WAS meant for. Rather, we should
argue what it should be meant for. Would you care to refute my
previously stated argument?
--
Eliot Lear
[le...@net.bio.net]
Peter da Silva
> Again, it is sort of useless to argue what the sys hierarchy WAS meant for.
Sounds like the old Jefferson/Hamilton arguments about the constitution.
Needless to say the net.aclu won't agree.