Re: 1st RFD: The Great Downsizing 2011/1
Steve Crook
Message-Id: <great-downsizi...@news.albasani.net>:
> - zero on-topic, non-crossposted threads in the past 18 months
> - on-topic questions that received no on-topic answer do not count
At the risk of subverting the thread, where does this info come from?
Alexander Bartolich
> Big-8 Management Board wrote:
> [...]
>> - zero on-topic, non-crossposted threads in the past 18 months
>> - on-topic questions that received no on-topic answer do not count
> At the risk of subverting the thread, where does this info come from?
I wrote a Perl script that queries the overview to determine the age
of oldest and newest article in a group. It is available here:
http://svn.schnuerpel.eu/viewvc.cgi/trunk/bin/get-retention.pl?root=schnuerpel&view=log
and can be retrieved like this:
svn co svn://svn.schnuerpel.eu/schnuerpel/trunk/bin/get-retention.pl
I asked a friend with an account with news.individual.de to run that
script. Apparently NID has a retention of 256 days for the BIG8. The
results was a list of 318 empty groups. 227 of them remained after
pruning the list for moderated groups, and names ending in ".misc".
On reader.albasani.net the retention for the BIG8 happens to be about 18
months. So I took the group list there, converted it into .newsrc format,
set up a score file that filters out all crossposts,
[*]
Score: =-9999
Newsgroups: ,
started slrn, loaded all groups, and had a look at them.
Apparently the overview data is not 100 % reliable. There were two
strange outliers, soc.culture.israel and sci.math.symbolic, which
are actually high traffic groups. But a lot of groups were really
empty, and most of them contained only a few posts which were not
that hard to check for on-topicness.
Ciao
Alexander.
Doug Freyburger
>
> This is a formal Request for Discussion (RFD) to remove the following
> 198 unmoderated newsgroups.
>
> - on-topic questions that received no on-topic answer do not count
Why were threads not messages chosen? To me posted questions indicate a
group is alive, lack of answers indicate a group is sick. I would have
trimmed groups based on on topic posts rather than threads.
> NEWSGROUP LINES:
>
> comp.lsi.cad Electrical Computer Aided Design.
Because I used to subscribe to this group I checked it. In the last
year it had a few questions that were not answered. As such it
addresses my question above.
Alexander Bartolich
> Big-8 Management Board wrote:
>>
>> This is a formal Request for Discussion (RFD) to remove the following
>> 198 unmoderated newsgroups.
>>
>> - on-topic questions that received no on-topic answer do not count
>
> Why were threads not messages chosen? To me posted questions indicate a
> group is alive, lack of answers indicate a group is sick. I would have
> trimmed groups based on on topic posts rather than threads.
Trimming groups is not an end in itself. The goal is to improve the ex-
perience of users by removing sources of frustration. Would you agree
that receiving no answers to on-topic questions is frustrating?
>> [...]
>> comp.lsi.cad Electrical Computer Aided Design.
>
> Because I used to subscribe to this group I checked it. In the last
> year it had a few questions that were not answered. As such it
> addresses my question above.
This is how the group looks like in my newsreader:
[alisha jon]: Ziphone: Latest Apple iPhone news!!!! [Wed, 26 Aug 2009 ]
[benoraj ]: IC printing company [Wed, 24 Mar 2010 ]
[shirin aal]: Problem about book: Algorithms for vlsi design au [Tue, 20 Apr 2010 ]
[Otto Haldi]: Witch CAD to draw simple Machanical parts? [Wed, 21 Apr 2010 ]
[comp.lsi.c]: 8mm tape drive to CDROM [Mon, 10 May 2010 ]
Ciao
Alexander.
Steve Bonine
> Trimming groups is not an end in itself. The goal is to improve the ex-
> perience of users by removing sources of frustration. Would you agree
> that receiving no answers to on-topic questions is frustrating?
It is an issue of whether a newsgroup that attracts the occasional
question, but is incapable of providing an answer, is sufficiently
functional to be left in the list. My personal opinion is, no. One of
the reasons I say this is that if the group isn't there maybe the
questions will be posted in a newsgroup where there are actually people
who can help.
I think that the bar has been set pretty low for this set of newsgroups,
and the fact that someone wandered by and asked a question doesn't
indicate to me that there's any life there. Of course one could
speculate that the very next time a question was posed that dozens of
people would offer helpful advice, but in the specific case of
comp.lsi.cad it appears that there were at least two on-topic questions
posed in the last year, neither of which prompted anything from a
potential group of lurkers.
One technique is to post directly in the newsgroup warning of its
potential death. This should elicit response from anyone who is
monitoring the newsgroup and believes that it still serves a purpose.
The problem with this technique is that an emotional tie to a newsgroup
does not constitute a reason for it to exist. Reading that list of
group names was like a waltz through history for me, but the fact that a
name brings back memories does not imply that the newsgroup serves any
purpose these days.
David E. Ross
The purpose of this RFD is to obtain feedback from the public. Such
feedback might expose situations in which a listed newsgroup should
indeed be retained and not eliminated.
If you see a newsgroup listed in the RFD that should be retained, please
reply to this RFD with a brief rationale for retention.
--
David E. Ross
<http://www.rossde.com/>
Anything I post in this newsgroup is my personal
opinion and does not reflect the official position
of the Big8-Usenet Board.
David Bostwick
[...]
>
>Trimming groups is not an end in itself. The goal is to improve the ex-
>perience of users by removing sources of frustration. Would you agree
>that receiving no answers to on-topic questions is frustrating?
>
To help relieve a little frustration from your life, I'll respond to this. :)
Trimming doesn't seem like a big deal to me. I don't think that trimming will
bring in users, but I'm not adamantly opposed to it. It just seems like much
ado about not a lot. Some questions have occurred to me, however.
How many reports of such frustration have been logged? How many people have
said they'd be more involved in Usenet if there were fewer empty groups? 6?
60?
Is the data anecdotal, or verifiable?
Are the supporting facts real, or assumed by those who want to trim groups?
How many rmgroups will be ignored by NSPs?
How will a potential user be directed to the active group for a topic? If
comp.lsi.cad is removed, for example, will there be another group that appears
when a search is made for CAD?
There may not be any concrete answers to those questions. I'm not sure the
answers matter, however, because the action appears to be already decided.
I'm far from a Usenet guru, but I don't see much return on investment. I'll
defer to those of you who are gurus, however. I'll stick around - 20 years of
checking news is a hard habit to break.
Theo Markettos
> Because of the magnitude of the group list this proposal is not cross-
> posted to target groups. In the course of these proceedings the B8MB
> will post pointers to this announcement to appropriate groups. Readers
> are encouraged to take initiative and spread the message.
Minor sidenote:
I just reposted the RFD in comp.sys.psion.* with an added explanation. But
this afterwards made me think... how vigorous are spam filters in this
respect? I added 10 lines of extra content; will that still count as
'excessive multi-posting' as the majority of the content is this the same?
How good are the comparison algorithms of things like cleanfeed in this
regard?
If other people do the same, how vulnerable is it for the original posting
to get wiped out by filters?
Theo
Alexander Bartolich
> In news.groups.proposals Big-8 Management Board <bo...@big-8.org> wrote:
>> Because of the magnitude of the group list this proposal is not cross-
>> posted to target groups. In the course of these proceedings the B8MB
>> will post pointers to this announcement to appropriate groups. Readers
>> are encouraged to take initiative and spread the message.
>
> Minor sidenote:
>
> I just reposted the RFD in comp.sys.psion.* with an added explanation.
The whole RFD, including the list of 200 groups? Whoa. I would have
just listed the two groups that the readers of comp.sys.psion.* might
be interested in. There is no need to let them wade through all of it.
comp.sys.psion.comm Discussions about Psion communications.
comp.sys.psion.marketplace Buy and sell Psion computers and accessories.
> But this afterwards made me think... how vigorous are spam filters in
> this respect? I added 10 lines of extra content; will that still count
> as 'excessive multi-posting' as the majority of the content is this the
> same? How good are the comparison algorithms of things like cleanfeed in
> this regard?
AFAIK Cleanfeed trims the whitespace and calculates the MD5 hash.
This catches only identical posts.
> If other people do the same, how vulnerable is it for the original
> posting to get wiped out by filters?
Cleanfeed can reject incoming messages based on the history of recently
received posts. Cleanfeed does not retro-actively remove posts that are
already stored in the news spool.
Ciao
Alexander.
Mark Kramer
Steve Bonine <s...@pobox.com> wrote:
>It is an issue of whether a newsgroup that attracts the occasional
>question, but is incapable of providing an answer, is sufficiently
>functional to be left in the list. My personal opinion is, no. One of
>the reasons I say this is that if the group isn't there maybe the
>questions will be posted in a newsgroup where there are actually people
>who can help.
Since the question is being asked in that group, it is assumed that it
is the best place to ask it.
If that group was the best place to ask it, it is not logical to assume
that people in some other group will be better able to answer it once
the assumed-dead group is removed.
"I've got an SGI O2 that I'm trying to process some real-time streaming
audio with. How do I..." wouldn't not likely find an answer specific to
SGI anywhere but comp.sys.sgi.audio.
>I think that the bar has been set pretty low for this set of newsgroups,
>and the fact that someone wandered by and asked a question doesn't
>indicate to me that there's any life there.
Likewise, the fact that nobody was able to answer the question, and
didn't think that posting a "good question" response was appropriate,
doesn't indicate the group is dead.
Dave Sill
> Likewise, the fact that nobody was able to answer the question, and
> didn't think that posting a "good question" response was appropriate,
> doesn't indicate the group is dead.
Right...it's just resting. <http://www.youtube.com/watch?v=npjOSLCR2hE>
But seriously... That by itself doesn't mean the group is dead, but the
fact that there's been no other discussion of anything for a year and
half...well, that *does*.
-Dave
Dave Sill
> Why were threads not messages chosen?
Because discussion groups without discussions are dead.
> To me posted questions indicate a
> group is alive,
No, at best they indicate that a group is undead. Maybe we *should* call
them zombie groups, but the point is that they're not alive, they serve
no useful purpose, they occasionally waste people's time, they clutter
news servers and group lists, and they just need to go away.
-Dave
Dave Sill
> Trimming doesn't seem like a big deal to me. I don't think that
> trimming will bring in users, but I'm not adamantly opposed to it. It
> just seems like much ado about not a lot.
Exactly, it's not a big deal. A few people will make a lot of noise about
it, but the bottom line is that nobody really cares if a bunch of dead
groups go away. It's just a little long-overdue housekeeping.
> How many reports of such frustration have been logged?
Logged? Let me check the Big 8 Customer Support Ticket Tracking
System...oops, there isn't one.
> Are the supporting facts real, or assumed by those who want to trim
> groups?
It's a fact that there are dead groups. We believe the target list for
this RFD consists entirely of dead groups. I've personally wasted time
posting to dead groups. I can't see any upside to keeping them around.
> How many rmgroups will be ignored by NSPs?
Who knows? Who cares? If a dead group isn't rmgroup'd on a poorly-run
server, and someone posts to it, they're probably not going to get a
response. True, the dead group won't propagate as well as if it hadn't
been rmgroup'd, but ensuring the propagation of dead groups is not a high
priority of the B8MB.
> How will a potential user be directed to the active group for a topic?
> If comp.lsi.cad is removed, for example, will there be another group
> that appears when a search is made for CAD?
Searching the newsgroups list with a newsreader or (horrors) Google
Groups reveals several possibilities including:
alt.cad
alt.cad.autocad
comp.cad.autocad
comp.cad.microstation
sci.electronics.cad
> There may not be any concrete answers to those questions. I'm not sure
> the answers matter, however, because the action appears to be already
> decided.
It's not a done deal until the rmgroups are issued, but I don't see any
showstoppers.
-Dave
Steve Bonine
> In article<8qqsej...@mid.individual.net>,
> Steve Bonine<s...@pobox.com> wrote:
>> It is an issue of whether a newsgroup that attracts the occasional
>> question, but is incapable of providing an answer, is sufficiently
>> functional to be left in the list. My personal opinion is, no. One of
>> the reasons I say this is that if the group isn't there maybe the
>> questions will be posted in a newsgroup where there are actually people
>> who can help.
>
> Since the question is being asked in that group, it is assumed that it
> is the best place to ask it.
This is the assumption made by the person who submitted the article. It
may indeed not be the best place, especially if there are many potential
newsgroups that are candidates.
> If that group was the best place to ask it, it is not logical to assume
> that people in some other group will be better able to answer it once
> the assumed-dead group is removed.
True, but you're basing your conclusion on a decision made by someone
who may not know the best place to ask.
> "I've got an SGI O2 that I'm trying to process some real-time streaming
> audio with. How do I..." wouldn't not likely find an answer specific to
> SGI anywhere but comp.sys.sgi.audio.
Not necessarily. If no one is reading comp.sys.sgi.audio, then
guaranteed you are not finding an answer there. If you put the question
into a more active newsgroup it will at least be seen by people who
might know the answer.
>> I think that the bar has been set pretty low for this set of newsgroups,
>> and the fact that someone wandered by and asked a question doesn't
>> indicate to me that there's any life there.
>
> Likewise, the fact that nobody was able to answer the question, and
> didn't think that posting a "good question" response was appropriate,
> doesn't indicate the group is dead.
No, it doesn't. But when you put that with the fact that there has been
no other activity in the newsgroup in a year, it's a good indication
that the newsgroup is serving no purpose.
I believe that removing the groups in the RFD from the newsgroups list
will have a net effect of improving the discussion in remaining
newsgroups. I'm not going to argue that the degree of improvement
justifies the effort to remove the groups, but I'm not the one putting
forth the effort.
Alexander Bartolich
> Alexander Bartolich <alexander...@gmx.at> wrote:
> [...]
>>Trimming groups is not an end in itself. The goal is to improve the ex-
>>perience of users by removing sources of frustration. Would you agree
>>that receiving no answers to on-topic questions is frustrating?
>
> To help relieve a little frustration from your life, I'll respond to this. :)
Thank you very much, kind Sir!
> [...]
> How many reports of such frustration have been logged? How many people have
> said they'd be more involved in Usenet if there were fewer empty groups? 6?
> 60?
>
> Is the data anecdotal, or verifiable?
>
> Are the supporting facts real, or assumed by those who want to trim groups?
Oh. The Wikipedian Protester. [1] This is an extremly strong move.
My only recourse is to start a meta discussion.
Debates on group removal are by their nature ideological. Because low
traffic numbers are undisputable facts the discussion typically centers
on the interpretation of those facts. Basically the argument is that
the cost of the storage required for empty groups is neglectable. Thus
even the smallest of benefits makes keeping those groups worthwhile.
Attempts to actually contemplate about infinitesimals are then drowned
in a sea of emotions.
You are right in that the trope of the frustrated user that reloads the
group again and again in rising agony is not the most scientific of ar-
guments. But then this is not a scientific debate.
> How many rmgroups will be ignored by NSPs?
Over 9000.
> How will a potential user be directed to the active group for a
> topic?
How does that happen today?
How did that happen before comp.lsi* was established?
Anyway, I think that it will be easier to search for active groups
when there is less distraction by inactive groups.
> If comp.lsi.cad is removed, for example, will there be another group
> that appears when a search is made for CAD?
comp.cad.* springs to mind.
> [...] I'll stick around - 20 years of checking news is a hard habit
> to break.
That is nice to hear.
Ciao
Alexander.
--
[1] http://xkcd.com/285/
--
Kathy Morgan
> If that group was the best place to ask it, it is not logical to assume
> that people in some other group will be better able to answer it once
> the assumed-dead group is removed.
That would be logical but often is not true. A question is more likely
to be answered in a group that other people are using, even if the
question is only peripherally related to the topic of the group.
--
Kathy
Kathy Morgan
> I just reposted the RFD in comp.sys.psion.* with an added explanation. But
> this afterwards made me think... how vigorous are spam filters in this
> respect?
The RFD is not posted to all the groups listed because that would by
definition be spam. Pointers that include all the orginal post plus a
small amount of additional material would also be considered to be spam;
if BI>20 then it is cancellable spam. I don't know how many spam
cancelers are still active or how good their algorithms are at catching
that sort of thing.
--
Kathy
Mark Kramer
Steve Bonine <s...@pobox.com> wrote:
>Mark Kramer wrote:
>> In article<8qqsej...@mid.individual.net>,
>> Steve Bonine<s...@pobox.com> wrote:
>>> It is an issue of whether a newsgroup that attracts the occasional
>>> question, but is incapable of providing an answer, is sufficiently
>>> functional to be left in the list. My personal opinion is, no. One of
>>> the reasons I say this is that if the group isn't there maybe the
>>> questions will be posted in a newsgroup where there are actually people
>>> who can help.
>>
>> Since the question is being asked in that group, it is assumed that it
>> is the best place to ask it.
>
>This is the assumption made by the person who submitted the article.
And it would be my assumption, too, since I assume the poster knows better
than I what his question is and what it deals with.
>It
>may indeed not be the best place, especially if there are many potential
>newsgroups that are candidates.
Then I would assume he would have cross-posted it to those several
groups. Unfortunately, that cross-posted question and other-group
answer would have been ignored in the traffic survey done to determine
the deadness of the group.
>> "I've got an SGI O2 that I'm trying to process some real-time streaming
>> audio with. How do I..." wouldn't not likely find an answer specific to
>> SGI anywhere but comp.sys.sgi.audio.
>
>Not necessarily. If no one is reading comp.sys.sgi.audio, then
>guaranteed you are not finding an answer there.
You have not counted how many people are reading that group. The only
thing that was counted was the number of non-crossposted 'on topic'
messages that got responses.
>If you put the question
>into a more active newsgroup it will at least be seen by people who
>might know the answer.
Getting it seen by more people who don't know the answer is just as
frustrating as it not being answered anywhere else. Moreso, because
those in the inappropriate group are more likley to flame the newbie
for posting in the wrong place. "But this is the closest to on-topic
group ever since the B8MB deleted the best group for this question..."
>No, it doesn't. But when you put that with the fact that there has been
>no other activity in the newsgroup in a year,
This is a fact not in evidence. Crossposted articles were not counted.
>I believe that removing the groups in the RFD from the newsgroups list
>will have a net effect of improving the discussion in remaining
>newsgroups.
How will the absence of groups that nobody is using going to do anything
to any of the other groups? Your assumption is that nobody posts or
reads these groups, so how can their absence possible have any effect
on anything?
Rob Kelk
<alexander...@gmx.at> wrote:
>Theo Markettos wrote:
<snip>
>> If other people do the same, how vulnerable is it for the original
>> posting to get wiped out by filters?
>
>Cleanfeed can reject incoming messages based on the history of recently
>received posts. Cleanfeed does not retro-actively remove posts that are
>already stored in the news spool.
And if Cleanfeed does start cancelling the reposts, people can simply
post pointers to the original thead instead.
Sample pointer:
Subject: There's a proposal to delete this group
There's some discussion over on news.groups.proposals about a proposal
to delete this group. Look for the thread titled "1st RFD: The Great
Downsizing 2011/1" for more information - the first post in that thread
is news:great-downsizi...@news.albasani.net
If everybody uses that exact sample pointer, then it'll be cancelled as
excessive multiposting, so change the wording appropriately. (You can
use more or less formal phrasing, you can add personal opinions, you can
and should rephrase the text so it sounds more like something you would
write, and so on.)
Mark Kramer
>Exactly, it's not a big deal. A few people will make a lot of noise about
>it, but the bottom line is that nobody really cares if a bunch of dead
>groups go away. It's just a little long-overdue housekeeping.
Usenet isn't just about the past or present. It is about the future. A group
that is silent today may have a use tomorrow. Given the complete lack of
consequences from NOT deleting it, why do it?
>It's a fact that there are dead groups. We believe the target list for
>this RFD consists entirely of dead groups. I've personally wasted time
>posting to dead groups. I can't see any upside to keeping them around.
I've personally spent time posting to dead groups, only to find that they
really weren't dead after all. I just had to ask a pertenent question that
someone could answer.
>> How many rmgroups will be ignored by NSPs?
>
>Who knows? Who cares?
People who think the damage from keeping an inactive group is so small
as to be completely ignorable, and the damage to Usenet from having
hundreds of tiny islands of cut-off discussion groups left over from
failed rmgroups...
>If a dead group isn't rmgroup'd on a poorly-run
>server, and someone posts to it, they're probably not going to get a
>response.
Thus they will be frustrated, and more frustrated than if the group had
some non-zero chance of reaching someone with an answer. The goal is
to remove frustration, not create it, wasn't it?
>True, the dead group won't propagate as well as if it hadn't
>been rmgroup'd, but ensuring the propagation of dead groups is not a high
>priority of the B8MB.
Circular reasoning.
Brad Templeton
readership. In those times there were lots of servers where
people read news on the server and had a .newsrc file. As such
many thousands of .newsrc files were used to gather statistics on who
was reading.
This is possible today, if somebody is ready to do modest technical
work, and gets permission from operators of some large NNTP servers
used for newsreading.
Step 1: Collect posting statistics. Generally disregard crossposts,
or at least those that didn't put the group first. Gather stats
on number of distinct users in From lines over various intervals of
time. You can also get stats on article sizes, presence of different
article forms (binaries, mime types, followups vs. originals, crossposts)
Step 2: Gather stats on injection servers, see which are most popular,
these are going to be the biggest reading servers.
Step 3: Get permission to analyse logs at injection major server in
privacy protecting way. Scan logs. Eliminate non-human readers
that, for example, fetch every article in a group (not just xover)
in quick succession, or who come at the exact same time each day.
For each group gather stats on total number of article reads,
number of unique article reads, median readership of typical article,
typical percentage of articles read by average or median readers
and so on.
I believe Google groups may post stats on how
Step 4: If you get enough servers can can map readership on that server
to posting over a large enough base, you can roughly extrapolate to all
of USENET to calculate readers. For example, if there are 10,000 readers
on your group of servers and those readers post 20% of the article, you
might guess there are 50K readers total.
You will have to only amalgamate and promise to never retain personal
data. And you won't get in on the large binary servers which promise not
to log reading.
But still this can be done, and it's work but not any more work than
the old arbitron was (probably less, in fact.)
This is the sort of data you need if you really want to do a measured
pruning of newsgroups.
Which I'm not sure you need to do, by the way, because another alternative
if you have this data is just to provide nice web sites where people can
search for newsgroups to read, and be told when they search "This is
probably a dead newsgroup" and why.
While sites do maintain a "newsgroups" file and an active file and newsreaders
do have means to browse and search it, truth is these suck, and you can
do this much better with a web interface. Using news: URLs, if they have
configured that in the browser and newsreader, lets you subscribe or browse
after searching.
I would even push for newsreaders that are still being updated to abandon
the use of the newsreader as a way to find groups. There does not have
to be only one index, there can be several, with different criteria of what
is dead.
--
Sail up the Yangtse river into China in my photojournals
http://www.templetons.com/brad/photo/china/
Winston
>> Trimming groups is not an end in itself. The goal is to improve the
to which david.b...@chemistry.gatech.edu (David Bostwick) replied:
> How many reports of such frustration have been logged?
WBE replies:
I would guess hardly any, but that guess is based only on my own
experience and reaction. I've used USENET for over 20 years, and
if I post a question that's not answered, I don't go looking for
some place to complain about how frustrating an experience that was.
:)
> How many people have said they'd be more involved in Usenet if there
> were fewer empty groups?
I don't recall anyone ever saying that, but I do recall a time or two when
it would have saved me a few days if I'd known to use group X instead of
(nearly-dead) group Y (whose name looked better). I got lucky, and someone
who occasionally read both groups told me of the better one.
> Is the data anecdotal, or verifiable?
My experience is anecdotal. YMMV. :)
As for the list of newsgroups itself, I read the entire list and recognized
many as ones I used to read years ago. None were groups I still read; some
are still in my .newsrc file as not-currently-subscribed groups (that I
haven't even looked at in years); and some refer to technologies like ISDN
that aren't much relevant these days. I don't plan to object to deletion
of any of the listed newsgroups.
Not that my lack of interest matters, since it's the people who *are*
interested a group that count...
OTOH, I don't think deleting ~200 groups out of the tens of thousands that
exist really qualifies to be called a "Great Downsizing". :)
-WBE
Winston
one that doesn't involve end-user statistics at all -- many smaller news
servers only carry groups their users have requested, or that look active.
Start by removing newsgroups they carry from the list of all newsgroups.
-WBE
Steve Bonine
> And if Cleanfeed does start cancelling the reposts, people can simply
> post pointers to the original thead instead.
>
> Sample pointer:
>
>
> Subject: There's a proposal to delete this group
>
> There's some discussion over on news.groups.proposals about a proposal
> to delete this group. Look for the thread titled "1st RFD: The Great
> Downsizing 2011/1" for more information - the first post in that thread
> is news:great-downsizi...@news.albasani.net
At the risk of incurring the ire of folks who think everything on Usenet
should be 100% Usenet-based, an option is to post the URL
http://www.big-8.org/wiki/Nan:2011-02-01-rfd-great-downsizing
which is a pointer to the text of the RFD at the big-8 site. Another URL is
http://groups.google.com/group/news.announce.newgroups/browse_thread/thread/26edeccb1d9c724e#
Brad Templeton
Yes, that's useful. Though for a low-volume group it is hard
to judge readership on, I presume. It's been a while since I
examined it. Do most newsreaders, when doing their first scan
to see new articles, do GROUP commands on all their groups, or
do they do LIST ACTIVE with patterns or other such commands?
In the old days you could pull down the whole active file but I
presume more often just a few are queried, since most people
subscribe to a tiny fraction of newsgroups. If so it becomes a
lot easier to measure stuff. It would be nice if NNTP had
added some sort of user-agent style field to be sent by
connecting software so you could keep stats on what tools
are reading with. If they post you sometimes see that but
again there are privacy consequences of logging this stuff.
Still, a lot can be learned and from that it can be figured
what groups are truly dead. Whether that's for a downsizing
(bad idea) or just a good searchable database that tells you
real information on the groups (better idea.)
--
What's the future of TV and advertising after Tivo?
http://www.templetons.com/brad/tvfuture.html
Rob Kelk
Those also work, and will help make the pointers distinct from each
other - and that will help them not be cancelled as substantially
identical posts. (I'm "lazy" and picked the one that was shortest to
type, is all.)
--
Rob Kelk Personal address (ROT-13): eboxryx -ng- tznvy -qbg- pbz
Any opinions here are mine, not ONAG's.
ott.* newsgroup charters: <http://onag.pinetree.org>
Any Usenet message claiming to be from me but posted from any server
other than individual.net is a forgery. Please filter out such
messages if you have the capability.
David Empson
> In the old days, Brian Reid published regular reports on newsgroup
> readership. In those times there were lots of servers where
> people read news on the server and had a .newsrc file. As such
> many thousands of .newsrc files were used to gather statistics on who
> was reading.
>
> This is possible today, if somebody is ready to do modest technical
> work, and gets permission from operators of some large NNTP servers
> used for newsreading.
>
> Step 1: Collect posting statistics. Generally disregard crossposts,
> or at least those that didn't put the group first. Gather stats
> on number of distinct users in From lines over various intervals of
> time. You can also get stats on article sizes, presence of different
> article forms (binaries, mime types, followups vs. originals, crossposts)
>
> Step 2: Gather stats on injection servers, see which are most popular,
> these are going to be the biggest reading servers.
>
> Step 3: Get permission to analyse logs at injection major server in
> privacy protecting way. Scan logs. Eliminate non-human readers
> that, for example, fetch every article in a group (not just xover)
> in quick succession, or who come at the exact same time each day.
If you eliminated readers that fetch every article in a group in quick
succession, that would ignore most people using an offline newsreader,
as I do. I normally fetch every new article in a group, and then
selectively read them at my leisure while disconnected from the server.
Followups or new articles are posted to the server on the next
connection.
Depending on the way I've configured my newsreader for a particular
group, it will either fetch all new messages in a single pass, or fetch
all new headers then the body of selected articles.
> For each group gather stats on total number of article reads,
> number of unique article reads, median readership of typical article,
> typical percentage of articles read by average or median readers
> and so on.
Also of limited use if a significant fraction of readers are using
offline newsreaders. My news server has no way of knowing which articles
I read (apart from when I post a followup), only which groups I read.
To take one platform-specific example, there are at least two quite
popular offline newsreaders on the Mac, so such a mechanism would
seriously distort statistics in Mac-specific groups.
> I believe Google groups may post stats on how
Did you intend to finish that sentence? :-)
--
David Empson
dem...@actrix.gen.nz
Martin X. Moleski, SJ
> ... I don't think deleting ~200 groups out of the tens of thousands that
>exist really qualifies to be called a "Great Downsizing". :)
The Big-8 isn't really that big, either. It's only
about 2400 groups. What's "great" about the "Great
Downsizing" is the decision to accelerate the removal
process by submitting the list of dead groups all at
once.
Of course, this will have no affect on alt.* or any
of the hundreds of hierarchies or tens of thousands
of groups not in the Big-8.
Marty
--
Member of the Big-8 Management Board (B8MB) <http://www.big-8.org>
Unless otherwise indicated, I speak for myself, not for the Board.
Alexander Bartolich
> [...]
> Step 3: Get permission to analyse logs at injection major server in
> privacy protecting way. Scan logs. Eliminate non-human readers
> that, for example, fetch every article in a group (not just xover)
> in quick succession, or who come at the exact same time each day.
>
> For each group gather stats on total number of article reads,
> number of unique article reads, median readership of typical article,
> typical percentage of articles read by average or median readers
> and so on.
Nowadays most news clients are so called off-line newsreaders. They
maintain a kind of private newsspool on the local hard disk. When a
user subscribes to a group they download about 100 of the most recent
articles, and then synchronize the with news server on every reconnect.
It is thus not possible to tell how many articles the user actually
reads. This information is available only to the newsreader.
> [...]
> You will have to only amalgamate and promise to never retain personal
> data. And you won't get in on the large binary servers which promise
> not to log reading.
In other words: Your idea does not work.
What might be possible is to extend Innreport with features similar to
ninpaths/sendinpaths. The problem here is that you can only pass on
anonymized accumulated data, so all the smart filtering you mention
needs to happen in Innreport. As that program is updated only infre-
quently it is very hard to push out updates. Getting a smart algorithm
full of heuristics right on the first attempt, however, is almost im-
possible.
Accumulating the raw reader statistics still would be interesting,
but ...
> But still this can be done, and it's work but not any more work than
> the old arbitron was (probably less, in fact.)
... your estimates on the effort required are way off mark.
> [...]
> This is the sort of data you need if you really want to do a measured
> pruning of newsgroups.
Users that frequently read all the non-existing posts of an empty
group can just as well read all the non-existing posts of a non-
existing group.
--
host -t mx moderators.isc.org
Doug Freyburger
>
> In the old days, Brian Reid published regular reports on newsgroup
> readership.
I remember those reports. Even then I understood something more basic
about UseNet: UseNet content is about the posters not about the
readers. For paid subscription services UseNet is about the number of
readers but they come for the content. Even when counting paid
subscriptions what counts is the traffic.
> Step 1: Collect posting statistics.
That's what was done and I suggest that's all that needs to be done.
There's a possible step 2 - Get news admins of paid services involved
and listen to their input. If that happens *everyone* will listen I
think. They are the ones who should care about the number of readers.
It's often discussed that their main revenue source is downloaded
binaries nottext only UseNet so it's unlikely any will post. If none
post I say all we need is post counts.
KalElFan
news:s4adncHLz5Q909TQ...@supernews.com...
> On Tue, 1 Feb 2011 20:42:38 CST, w...@ubeblock.psr.com.invalid (Winston)
> wrote in <ydtygn4...@UBEblock.psr.com>:
>
>> ... I don't think deleting ~200 groups out of the tens of thousands
>> that exist really qualifies to be called a "Great Downsizing". :)
>
> The Big-8 isn't really that big, either. It's only
> about 2400 groups.
A while back I looked at the Board's Checkgroups lists and did a
very rough line count and came up with a bit less than that. Just
re-checking it again now, the List of Big 8 Newsgroups posting from
January 15, 2011 to news.announce.newgroups has 2238 lines per
the header, and I count about 35 lines in total before and after the
group list. One line per group (I didn't check every one to make
sure but presumably that 1:1 relationship holds) would mean 2203
groups on the list. So say 2200. Is the 2,400 number including some
not on the current list, previously removed or whatnot and so 2,400
was the peak number but not now?
> What's "great" about the "Great Downsizing" is the decision to
> accelerate the removal process by submitting the list of dead groups
> all at once.
Does the 2011/1 at the end of the subject line suggest the Downsizing
net will be widened for a 2011/2, and so on pass though? Or is it the
Board's intent that this will be it for 2011 removals? Or is the "1" just
a "1st" RFD reference repeated at the end of the Suject Line?
> Of course, this will have no affect on alt.* or any of the hundreds
> of hierarchies or tens of thousands of groups not in the Big-8.
Right, and those other groups are all part of Usenet as well. Winston
may or may not have known that the Big 8 has only the 2000+ groups
and not the tens of thousands he alluded to, but he can let us know
if he wants and my assumption is he knew the Big 8 had only a fraction
of the groups.
Anyway, there are many other issues I'll probably comment on but that's
it for now. So to summarize, some clarification on the 2400 vs. 2200,
and on the "2011/1" and what that means, and on the Board's intent
BEYOND this less than 10% culling of what they consider groups that
should be removed.
Martin X. Moleski, SJ
> ... Is the 2,400 number including some
>not on the current list, previously removed or whatnot and so 2,400
>was the peak number but not now?
Yes. I was working from memory, not from any current list.
>Does the 2011/1 at the end of the subject line suggest the Downsizing
>net will be widened for a 2011/2, and so on pass though? Or is it the
>Board's intent that this will be it for 2011 removals? Or is the "1" just
>a "1st" RFD reference repeated at the end of the Suject Line?
I think it means "Part I."
>... some clarification on the 2400 vs. 2200,
>and on the "2011/1" and what that means, and on the Board's intent
>BEYOND this less than 10% culling of what they consider groups that
>should be removed.
That last phrase seems a little tangled.
The "Great Downsizing 2011/1" culls less than 10% of the
groups that exist in the Big-8.
It is not 10% of the groups that we think should be removed.
I'm not sure what fraction it will turn out to be when all
is said and done.
Part II, if there is a Part II, will be worked out after
we show how Part I goes.
Charles Lindsey
> REQUEST FOR DISCUSSION (RFD)
>This is a formal Request for Discussion (RFD) to remove the following
>198 unmoderated newsgroups.
>news.admin.nocem NoCeM protocol policy issues and information.
That's the only one that bothers me. I don't subscribe to it, but I like
the Nocem protocol and wish it were more widely recognized and used.
>sci.aquaria Only scientifically-oriented postings about aquaria.
Ah! So the aquariums finally won :-) .
--
Charles H. Lindsey ---------At Home, doing my own thing------------------------
Tel: +44 161 436 6131 Web: http://www.cs.man.ac.uk/~chl
Email: c...@clerew.man.ac.uk Snail: 5 Clerewood Ave, CHEADLE, SK8 3JU, U.K.
PGP: 2C15F1A9 Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5
Brad Templeton
Doug Freyburger <dfre...@yahoo.com> wrote:
>Brad Templeton wrote:
>>
>> In the old days, Brian Reid published regular reports on newsgroup
>> readership.
>
>I remember those reports. Even then I understood something more basic
>about UseNet: UseNet content is about the posters not about the
>readers. For paid subscription services UseNet is about the number of
>readers but they come for the content. Even when counting paid
>subscriptions what counts is the traffic.
>
>> Step 1: Collect posting statistics.
>
>That's what was done and I suggest that's all that needs to be done.
>
No, what was calculated back then was both posting stats and readership
stats -- how many users on a site had the group in their .newsrc and
had current articles within it marked as read, I think.
--
Travel the coast of Oregon in my photojournals
http://www.templetons.com/brad/photo/oregon/
Brad Templeton
David Empson <dem...@actrix.gen.nz> wrote:
>Brad Templeton <b...@templetons.com> wrote:
>If you eliminated readers that fetch every article in a group in quick
>succession, that would ignore most people using an offline newsreader,
>as I do. I normally fetch every new article in a group, and then
>selectively read them at my leisure while disconnected from the server.
>Followups or new articles are posted to the server on the next
>connection.
Yes, it would ignore such people. Well, it would be possible to
detect their patterns and count them, but not how many articles they
read, or if they are still reading (or just sucking down articles but
skipping the group.)
I am sure you could learn something.
>
>To take one platform-specific example, there are at least two quite
>popular offline newsreaders on the Mac, so such a mechanism would
>seriously distort statistics in Mac-specific groups.
Perhaps. As noted, you could detect their presence at least, if
not analyse reading habits, so this could help. It's adequate for
at least spotting dead newsgroups.
And of course, as has often been done to measure web readership,
many programs have existed where people opted in to statistical
examination of their client activity (alexa, most toolbars) so the
offline readers can be measured if their are maintained and their
authors want this to be the case.
>
>> I believe Google groups may post stats on how
>
>Did you intend to finish that sentence? :-)
Sorry. Google groups does post stats on posting volumes and
web readership volumes of newsgroups. That's a start there, if
it can be scraped, or if the Google Groups team is willing to offer
up the data in a file. I could try to ask them.
Winston
>>> ... I don't think deleting ~200 groups out of the tens of thousands
>>> that exist really qualifies to be called a "Great Downsizing". :)
"Martin X. Moleski, SJ" replied:
>> The Big-8 isn't really that big, either. It's only
>> about 2400 groups.
[...]
>> What's "great" about the "Great Downsizing" is the decision to
>> accelerate the removal process by submitting the list of dead groups
>> all at once.
Um, okay. :) That being the case, I probably would have used a less
attention-grabbing subject, such as
"Proposed Big-8 Newsgroup Removals 2011#1".
>> Of course, this will have no affect on alt.* or any of the hundreds
>> of hierarchies or tens of thousands of groups not in the Big-8.
to which "KalElFan" <kale...@yanospamhoo.com> replied:
> Right, and those other groups are all part of Usenet as well. Winston
> may or may not have known that the Big 8 has only the 2000+ groups
> and not the tens of thousands he alluded to,
You're right, my mistake. Not only did I have the impression that the
number of Big-8 groups was larger, but when I made the earlier comment, I
was thinking of the complete set of all USENET newsgroups (the kind of
numbers USENET service providers sometimes advertise), so instead of a 9%
downsizing, removing 200 groups would be just a ~0.2% downsizing.
-WBE
Mark Kramer
Brad Templeton <b...@templetons.com> wrote:
>Yes, it would ignore such people. Well, it would be possible to
>detect their patterns and count them, but not how many articles they
>read, or if they are still reading (or just sucking down articles but
>skipping the group.)
>
>I am sure you could learn something.
And how would you count the readers who are on the distal end of a
"suck" feed? Or reading at the destination end of a news to mail
gateway?
Mark Kramer
There is a much simpler method: the user can figure it out for himself,
using his own criteria.
If a new user goes to a new-to-him group and sees NOTHING, it is pretty
obvious there is nothing there to see and the group is probably dead. It
doesn't take a world-side survey of major news servers.
What this proposal will result in is increased frustration for existing
users. Those people who we want to stay here because they will be the
anchors for any newbies we can manage to draw here.
How can it not be frustrating for someone who monitors a newsgroup on
a regular basis, ready to respond to any discussion that comes up he's
interested in, to come by one day and be told by his newsreader that
"that group doesn't exist"? Is the server broken? After spending some
time trying to find out what happened, he'll find out that his group was
deleted by people who don't read it and didn't bother letting him know
that they were even thinking about deleting it.
If that user says "hey, can you put it back, he'll be faced with an opaque
process that will require him to beat the bushes to get other supporters
and then incur a requirement to contact NSPs around the world to try to
get them to add it back. For a group that went through all that hassle
once already.
Why bother? He can go the web and either find a place already there,
or create one of his own easier.
Why not just keep from frustrating people and leave the groups that
aren't causing any harm? Especially when at least one person has just
reported HIS anecdotal evidence that posting to a dead group got him
a response telling him a better place -- not the answer itself, but a
dead group that really wasn't.
As an aside, the idea that "automated" reading shouldn't be counted in
"reader metrics", which others have already shown to be invalid, is
invalid for other reasons. Automated processing is the heart of Usenet,
and "suck" feeds used to be common. And, of course, I have processing
that scans the groups I want automatically so I won't miss anything if
I'm out for a week or two.
Winston
Or, on a considerably less serious note, since 9% is close to 10%, I
suppose it could have been titled "RFD: Decimation of Big-8 Newsgroups".
:-)
-WBE
Steve Bonine
> Doug Freyburger wrote:
>> Brad Templeton wrote:
>>>
>>> In the old days, Brian Reid published regular reports on newsgroup
>>> readership.
>>> Step 1: Collect posting statistics.
>>
>> That's what was done and I suggest that's all that needs to be done.
>>
> No, what was calculated back then was both posting stats and readership
> stats -- how many users on a site had the group in their .newsrc and
> had current articles within it marked as read, I think.
I believe that Doug's comment was based on what was done to determine
which groups to list in the RFD, not Brian Reid's statistics.
If we were trying to identify "slow" groups, I'd be interested in
obtaining better statistics. But for these 200 newsgroups, the Monty
Python parrot skit applies . . . they're dead. If the process continues
to the point that groups with measurable activity are being considered,
methods to measure it will be needed.
In the interval since the RFD was posted, not a single person has
identified a single newsgroup they think should be kept, with the
exception of the NoCeM group and that wish was based on a desire that
NoCeM's were more widely used. As much as I wish otherwise, keeping the
newsgroup is unlikely to have any effect on use of the protocol.
Doug Freyburger
> Doug Freyburger <dfre...@yahoo.com> wrote:
>>Brad Templeton wrote:
>>>
>>> In the old days, Brian Reid published regular reports on newsgroup
>>> readership.
The ancient arbitron posts.
>>I remember those reports. Even then I understood something more basic
>>about UseNet: UseNet content is about the posters not about the
>>readers. For paid subscription services UseNet is about the number of
>>readers but they come for the content. Even when counting paid
>>subscriptions what counts is the traffic.
>
>>> Step 1: Collect posting statistics.
>>
>>That's what was done and I suggest that's all that needs to be done.
The modern Alexander B post.
> No, what was calculated back then was both posting stats and readership
> stats -- how many users on a site had the group in their .newsrc and
> had current articles within it marked as read, I think.
I recall reading the ancient arbitron posts by Brian Reid over a 2400
baud modem connection as ihnp4!escher!doug. Your description of how
they were calculated matches my recollection of that. You don't agree
with how the modern calculation was used. I get that. I'm not sure I
need to care that we disagree on the topic.
I claim that what matters on Usenet is the posts and thus the posters.
The lurkers do not matter. The result of my claim is a calcuation that
was easy to make. Alexander B made that calculation.
You claim that what matters on UseNet is the readings and therefore the
readers. The lurkers are what matter. The result of your claim is a
calculation that is difficult to make. Actually it should be easier to
tell how many subscribe to a group on a server - Check the logs for how
many times a group was checked for new headers. I subscribe to plenty
of groups I have not posted to in years.
Why should a claim that is difficult to gather data for be better than a
claim that is easy to gather data for? There is "the lamppost effect"
to be considered. I take it you think that's what is happening here?
Better data rules as far as I am concerned. If you can show that one of
the groups on the list has thousands of lurking subscribers I will start
to ponder if I should join you in asking to keep it. But I wonder at
the value of it. Does it really matter if a group has tens of thousands
of silent lurkers? UseNet is a content network and lurkers do not
contribute content.
D. Stussy
news:great-downsizi...@news.albasani.net...
> This is a formal Request for Discussion (RFD) to remove the following
>
> Because of the magnitude of the group list this proposal is not cross-
> posted to target groups. In the course of these proceedings the B8MB
> will post pointers to this announcement to appropriate groups. Readers
> are encouraged to take initiative and spread the message.
I think that in this case, multi-posting (cross-posting 5 groups at a time)
may be appropriate.
I have comments on only 3 groups:
> comp.std.internat Discussion about international standards.
Consider that for "comp.std.internat", some may have avoided the group
thinking that it is a misspelled group - "comp.std.internEt"
(capitalization for emphasis). Therefore, before outright removal,
consider changing its name to "comp.std.international" and see if that
brings any life to the topic.
> news.admin.nocem NoCeM protocol policy issues and information.
As NoCeM is in use via "news.lists.filters", I would consider
"news.admin.nocem" as potentially Usenet infrastructure critical, and
therefore, it should not be removed.
> rec.music.artists.stevie-nicks Singer/songwriter Stevie Nicks.
As she is a currently touring artist (she's having a concert in Los Angeles
with Rod Stewart in May 2011), perhaps this choice should be reconsidered.
Dave Sill
>> comp.std.internat Discussion about international standards.
>
> Consider that for "comp.std.internat", some may have avoided the group
> thinking that it is a misspelled group - "comp.std.internEt"
> (capitalization for emphasis). Therefore, before outright removal,
> consider changing its name to "comp.std.international" and see if that
> brings any life to the topic.
Or we could nuke it 'cause it's dead. If someone wants to propose
comp.std.international, I'd have no objection.
>> news.admin.nocem NoCeM protocol policy issues and information.
>
> As NoCeM is in use via "news.lists.filters", I would consider
> "news.admin.nocem" as potentially Usenet infrastructure critical, and
> therefore, it should not be removed.
Critical but unused infrastructure? There are plenty of news.* groups in
which NoCeM could still be discussed.
>> rec.music.artists.stevie-nicks Singer/songwriter Stevie Nicks.
>
> As she is a currently touring artist (she's having a concert in Los
> Angeles with Rod Stewart in May 2011), perhaps this choice should be
> reconsidered.
Currently touring but the group is dead? Nuke it.
Keeping dead groups because they might wake up someday is even less
likely to succeed than the Field of Dreams approach.
-Dave
Alexander Bartolich
> "Martin X. Moleski, SJ" wrote:
>> w...@ubeblock.psr.com.invalid (Winston) wrote:
>>
>>> ... I don't think deleting ~200 groups out of the tens of thousands
>>> that exist really qualifies to be called a "Great Downsizing". :)
>>
>> The Big-8 isn't really that big, either. It's only
>> about 2400 groups.
>
> A while back I looked at the Board's Checkgroups lists and did a
> So say 2200.
Not a bad estimate. As of today there are 2216 groups.
> [...]
> Does the 2011/1 at the end of the subject line suggest the Downsizing
> net will be widened for a 2011/2, and so on pass though?
That's my intention.
What I know is that about 1100 unmoderated groups saw less than one post
per day in the past 18 months. And I guess that some of the groups with
sufficient traffic are just full of spam and kook sewage.
>> [...]
>> Of course, this will have no affect on alt.* or any of the hundreds
>> of hierarchies or tens of thousands of groups not in the Big-8.
>
> Right, and those other groups are all part of Usenet as well. Winston
> may or may not have known that the Big 8 has only the 2000+ groups
> and not the tens of thousands he alluded to, but he can let us know
> if he wants and my assumption is he knew the Big 8 had only a fraction
> of the groups.
The grand vision is to establish the BIG8 as a quality brand name.
Users browsing our group list should find only good (TM) groups there,
vibrant in activity and a pleasure to read.
I also expect the number of hobbiest news servers to rise, and for the
operators of such sites the utilization of hierarchies is important.
> [...]
> Anyway, there are many other issues I'll probably comment on but that's
> it for now. So to summarize, some clarification on the 2400 vs. 2200,
> and on the "2011/1" and what that means, and on the Board's intent
> BEYOND this less than 10% culling of what they consider groups that
> should be removed.
The Great Downsizing has the potential to cull about half of the groups
in the BIG8. Opinions will diverge on whether that's really such a great
idea. But the extent of the operation certainly qualifies it to be named
in memory of the Great Renaming.
Ciao
Alexander.
David Bostwick
[...]
>
>Why not just keep from frustrating people and leave the groups that
>aren't causing any harm? Especially when at least one person has just
>reported HIS anecdotal evidence that posting to a dead group got him
>a response telling him a better place -- not the answer itself, but a
>dead group that really wasn't.
I think I asked this question during the last round of rmgroups, or perhaps
when cleaning things up was first discussed quite a while ago. What is the
driving force behind removing groups? It's not like cleaning out your attic
to make more room for other stuff. Electrons or magnetic spots don't take up
a lot of space.
I don't recall ever getting a real answer, except for the need to organize
things better. If I want to know about CAD, I'll search for it. If I find
empty groups, so be it, but I may find what I need. I don't see how seeing a
few empty groups in my search is frustrating, especially if I find an active
one. If all the groups about CAD are empty, the ultimate result won't be
different whether the empty groups are removed or if they're still in the
list. Is it better, however, to see no groups, or to have a list that might
contain a "dead" group that turns out to be useful after all?
I think, however, that this has become A Mission From God, and will occur no
matter what. IOU.
D. Stussy
news:8qu1ju...@mid.individual.net...
> On Wed, 02 Feb 2011 16:32:30 -0600, D. Stussy wrote:
> >> comp.std.internat Discussion about international standards.
> >
> > Consider that for "comp.std.internat", some may have avoided the group
> > thinking that it is a misspelled group - "comp.std.internEt"
> > (capitalization for emphasis). Therefore, before outright removal,
> > consider changing its name to "comp.std.international" and see if that
> > brings any life to the topic.
>
> Or we could nuke it 'cause it's dead. If someone wants to propose
> comp.std.international, I'd have no objection.
Technically, I don't really care about this group. I never read it.
However, we all were asked for any reason (not to remove it), and I pointed
out one (to rename it).
> >> news.admin.nocem NoCeM protocol policy issues and information.
> >
> > As NoCeM is in use via "news.lists.filters", I would consider
> > "news.admin.nocem" as potentially Usenet infrastructure critical, and
> > therefore, it should not be removed.
>
> Critical but unused infrastructure? There are plenty of news.* groups in
> which NoCeM could still be discussed.
If it's critical, it doesn't matter if it's currently unused. It has a
purpose whereby it should remain for the smooth running of Usenet.
> >> rec.music.artists.stevie-nicks Singer/songwriter Stevie Nicks.
> >
> > As she is a currently touring artist (she's having a concert in Los
> > Angeles with Rod Stewart in May 2011), perhaps this choice should be
> > reconsidered.
>
> Currently touring but the group is dead? Nuke it.
What I am saying is:
1) There's a potential for new discussion based on her recent tour.
2) I note that there is no other group, even under alt, that deals with
her as a solo artist. (However, there exists "alt.music.fleetwood-mac".)
> Keeping dead groups because they might wake up someday is even less
> likely to succeed than the Field of Dreams approach.
Considering that many of the groups proposed for the chopping block are for
things that will NEVER return, this situation appears different. I say:
Let the group remain for now on a 6 month reprieve (skipping this removal
cycle), but if there still isn't activity after her concert tour, then
remove it.
Ont of 198 groups, I have no problem with removal for 195 of them. That's
my opinion. Deal with it.
Winston
> The Great Downsizing has the potential to cull about half of the groups
> in the BIG8.
Wow. OK. That wasn't obvious from the initial message.
In that case, the name does sound pretty appropriate.
I'm a little surprised that that many Big-8 groups would be useless.
-WBE
Winston
general point...
Doug Freyburger <dfre...@yahoo.com> writes:
> If you can show that one of the groups on the list has thousands of
> lurking subscribers I will start to ponder if I should join you in asking
> to keep it. But I wonder at the value of it. Does it really matter if a
> group has tens of thousands of silent lurkers?
* It might. Don't forget cases like rec.humor.funny where the lack of
posts was a moderation problem, not lack of readers. I think it
entirely possible that that group had tens of thousands of silent
lurkers. I followed comp.security.announce long after CERT stopped
posting to it, and even emailed them asking about the lack of posts.
You, at least, got an answer. I never did.
* Even with unmoderated groups, I can imagine a situation such as
alt.fan.la-femme.nikita which had mostly died because the show had ended
many years ago, but then CW came out with _Nikita_ and the group
revived. OTOH, newsgroups about ISDN and Amiga OS aren't likely to
revive once they're dead.
However, as I said, I didn't spot any groups in the list that looked worth
keeping if they're currently dead.
-WBE
Alexander Bartolich
>> > As NoCeM is in use via "news.lists.filters", I would consider
>> > "news.admin.nocem" as potentially Usenet infrastructure critical, and
>> > therefore, it should not be removed.
>>
>> Critical but unused infrastructure? There are plenty of news.* groups in
>> which NoCeM could still be discussed.
>
> If it's critical, it doesn't matter if it's currently unused. It has a
> purpose whereby it should remain for the smooth running of Usenet.
News.lists.filters is the only piece of infrastructure required to use
NoCeMs. Questions about NoCeM should be asked in news.software.nntp.
News.admin.nocem is just an annoying distraction.
Ciao
Alexander.
Martin X. Moleski, SJ
<04qdnbXCv-6zT9TQ...@earthlink.com>:
> ... What is the
> driving force behind removing groups? ...
It may be genetic. I don't think that I could give an answer
that would cover all the bases.
It's not economics.
It's not because of technical difficulties with carrying lots
of empty groups.
I would speculate that it is largely a matter of aesthetics.
The idea of group removals made its way into the mission statement
for the board back in 2006:
The Big-8 Management Board:
- creates well-named, well-used newsgroups in the
Big-8 Usenet hierarchies;
- makes necessary adjustments to existing groups;
- removes groups that are not well-used; and
- assists and encourages the support of a canonical
Big-8 newsgroup list by Usenet sites.
I do not have a strong drive either way. I have never led
a charge to remove groups; I've also never stood in the way
of those who are eager to do so.
I will probably vote for the removal of most of the 198 groups
on the list. I see no harm in dropping a few from the list
per OPs' suggestions.
>I think, however, that this has become A Mission From God,
>and will occur no matter what.
I agree that it is almost certainly going to occur.
I guess we'll find out whether it makes any difference
after it's done. "One observation is worth 10,000
expert opinions." ;o)
Rob Kelk
wrote:
And why does it matter? We're talking about groups that have had no
posts in the last 18 months - there's nothing to _be_ grabbed by a
"suck" feed (except possibly spam).
A group could have six billion readers for all we know, but if nobody's
posting to it, it's dead.
Rob Kelk
wrote:
>In article <ydpqrb4...@UBEblock.psr.com>,
>Winston <w...@ubeblock.psr.com.invalid> wrote:
>> If one only cares about identifying "dead" groups, another approach is
>>one that doesn't involve end-user statistics at all -- many smaller news
>>servers only carry groups their users have requested, or that look active.
>>Start by removing newsgroups they carry from the list of all newsgroups.
>
>There is a much simpler method: the user can figure it out for himself,
>using his own criteria.
Here's a criterion: "If I'm not subscribed to the group, it's a waste
of space." Since I'm a user of Usenet, if you're accepting the users'
criteria, when are we getting rid of the 2207 excess Big-8 groups and
leaving behind the 9 that are useful? Leaving this up to the user is
not desirable, IMHO.
If we use objective criteria instead of leaving it up to each user, we
can justify any group removal carried out by pointing to the objective
criteria and saying the removed groups failed to meet them.
The question here is what those objective criteria should be. "No
replies to on-topic posts in 18 months" seems a reasonable criterion to
me.
>If a new user goes to a new-to-him group and sees NOTHING, it is pretty
>obvious there is nothing there to see and the group is probably dead. It
>doesn't take a world-side survey of major news servers.
>
>What this proposal will result in is increased frustration for existing
>users. Those people who we want to stay here because they will be the
>anchors for any newbies we can manage to draw here.
Why? Currently, "a new user goes to a new-to-him group and sees
NOTHING". After the group is removed, the new user won't be able to go
to that group, and thus won't become frustrated in the first place.
>How can it not be frustrating for someone who monitors a newsgroup on
>a regular basis, ready to respond to any discussion that comes up he's
>interested in, to come by one day and be told by his newsreader that
>"that group doesn't exist"?
If there are no posters in a group, why is anyone lurking there? If
there are lurkers "ready to respond," why aren't they posting? Do they
want the group to die from lack of activity?
<snip>
>As an aside, the idea that "automated" reading shouldn't be counted in
>"reader metrics", which others have already shown to be invalid, is
>invalid for other reasons. Automated processing is the heart of Usenet,
>and "suck" feeds used to be common. And, of course, I have processing
>that scans the groups I want automatically so I won't miss anything if
>I'm out for a week or two.
It doesn't matter what is or isn't included in reader metrics, because
reader metrics are not an indicator of the health of a group. It
doesn't matter how many people are ready to read something in a group if
there's nothing there _to_ read.
What's important in the determination of whether a group is healthy is
poster metrics, not reader metrics. How often are people (or a person)
using the group to communicate to others, and how much of that
communication is appropriate for the group (for whatever definition of
"appropriate for the group" one wishes to use)? Anything else is
irrelevant, IMHO.
Rob Kelk
<spam+ne...@bde-arc.ampr.org> wrote:
>"Dave Sill" <da...@sill.org> wrote in message
>news:8qu1ju...@mid.individual.net...
>> On Wed, 02 Feb 2011 16:32:30 -0600, D. Stussy wrote:
>> >> comp.std.internat Discussion about international standards.
>> >
>> > Consider that for "comp.std.internat", some may have avoided the group
>> > thinking that it is a misspelled group - "comp.std.internEt"
>> > (capitalization for emphasis). Therefore, before outright removal,
>> > consider changing its name to "comp.std.international" and see if that
>> > brings any life to the topic.
>>
>> Or we could nuke it 'cause it's dead. If someone wants to propose
>> comp.std.international, I'd have no objection.
>
>Technically, I don't really care about this group. I never read it.
>However, we all were asked for any reason (not to remove it), and I pointed
>out one (to rename it).
The creation of that group was before my time, so I'm not certain, but I
believe the name was chosen because "internat" is eight characters long.
Older computers couldn't handle name elements longer than eight
characters. (Messages for group "comp.std.internat" would be stored in
directory "C:\spool\comp\std\internat\" or something similar, so the
"ional" part of the "international" had to be dropped to prevent the
computer complaining.) Is anyone still running antique software on a
production news system? I hope not, but I don't know for sure...
>> >> news.admin.nocem NoCeM protocol policy issues and information.
>> >
>> > As NoCeM is in use via "news.lists.filters", I would consider
>> > "news.admin.nocem" as potentially Usenet infrastructure critical, and
>> > therefore, it should not be removed.
>>
>> Critical but unused infrastructure? There are plenty of news.* groups in
>> which NoCeM could still be discussed.
>
>If it's critical, it doesn't matter if it's currently unused. It has a
>purpose whereby it should remain for the smooth running of Usenet.
Usenet appears to be running smoothly even without this group being
active.
>> >> rec.music.artists.stevie-nicks Singer/songwriter Stevie Nicks.
>> >
>> > As she is a currently touring artist (she's having a concert in Los
>> > Angeles with Rod Stewart in May 2011), perhaps this choice should be
>> > reconsidered.
>>
>> Currently touring but the group is dead? Nuke it.
>
>What I am saying is:
>1) There's a potential for new discussion based on her recent tour.
There's potential for anything. Is that potential being turned into
actuality?
>2) I note that there is no other group, even under alt, that deals with
>her as a solo artist. (However, there exists "alt.music.fleetwood-mac".)
I don't see this as a compelling reason to keep the group - the Big-8
isn't a vanity press, after all. (The closest Usenet has to that is the
alt.fan.* subhierarchy.)
>> Keeping dead groups because they might wake up someday is even less
>> likely to succeed than the Field of Dreams approach.
>
>Considering that many of the groups proposed for the chopping block are for
>things that will NEVER return, this situation appears different. I say:
>Let the group remain for now on a 6 month reprieve (skipping this removal
>cycle), but if there still isn't activity after her concert tour, then
>remove it.
>
>
>Ont of 198 groups, I have no problem with removal for 195 of them. That's
>my opinion. Deal with it.
Out of 198 groups, I have no problem with removal of all 198. That's my
opinion. Deal with it.
--
Steve Bonine
> I think I asked this question during the last round of rmgroups, or perhaps
> when cleaning things up was first discussed quite a while ago. What is the
> driving force behind removing groups? It's not like cleaning out your attic
> to make more room for other stuff. Electrons or magnetic spots don't take up
> a lot of space.
I'd like to turn this around. Do you have an objection to removing dead
groups from the list? If someone else wants to identify newsgroups with
no activity and prune them from the canonical list, is there a downside
to that? It might even improve the quality of discussion in the
remaining ones.
To me, a managed hierarchy has only one asset: the canonical list of
newsgroups. The flip side of adding of newsgroups is removing them.
Removal of dead groups was neglected so long that the big-8 list became
a list of groups that were approved by whatever mechanism was in place
at that particular moment in history, rather than a list of working
newsgroups.
The board has decided to invest time in removing nonfunctional
newsgroups from the list. Based on my belief that it should be a living
document, I support that effort. This is especially easy because my
time investment is almost zero. I sit back and say, "Go for it."
So if there's no downside, and potential improvement from doing it, why not?
Martin X. Moleski, SJ
> ... I
>believe the name was chosen because "internat" is eight characters long.
>Older computers couldn't handle name elements longer than eight
>characters. (Messages for group "comp.std.internat" would be stored in
>directory "C:\spool\comp\std\internat\" or something similar, so the
>"ional" part of the "international" had to be dropped to prevent the
>computer complaining.) Is anyone still running antique software on a
>production news system? I hope not, but I don't know for sure...
The limit of 8 characters per component of the name was dropped
over the years. There are many newsgroups now with longer
components (e.g., news.groups.proposals). For more details,
see:
http://moleski.net/newsgroups/b8_names.htm#r2_7
Kathy Morgan
> Doug Freyburger <dfre...@yahoo.com> writes:
> > If you can show that one of the groups on the list has thousands of
> > lurking subscribers I will start to ponder if I should join you in asking
> > to keep it. But I wonder at the value of it. Does it really matter if a
> > group has tens of thousands of silent lurkers?
>
> * It might. Don't forget cases like rec.humor.funny where the lack of
> posts was a moderation problem, not lack of readers.
That's part of the reason moderated groups were excluded from the
current round of downsizing.
--
Kathy
Harald Maedl
> The grand vision is to establish the BIG8 as a quality brand name.
"A man with visions should visit a doctor" [1]
"or should have the power to make it real[2],
but "Never trust doctors!" [3]
[1] H.Schmidt, German ex-chancellor
[2] H.Maedl, German ex-d.a.n.g-netizen
[3] Dr. House, German ex-philanthrope
> Users browsing our group list should find only good (TM) groups there,
> vibrant in activity and a pleasure to read.
For good and interesting newsgroups you would need netizens who are
bestirring for good and interesting news.
The only result of a consolidation of newsgroups could be, that you can
scrape users and themes together to built up bigger regular structures
and a better theme interleaving.
But there is no automatism between the consolidation of newsgroups and
the promised high-quality content.
"Quality rather than quantity" may be a credo, but I'm not convinced
that it will be possible only by removing some groups. May be it could
improve the conditions for further actions.
> I also expect the number of hobbiest news servers to rise, and for the
> operators of such sites the utilization of hierarchies is important.
Yezzz, famously with the very simple rookie-software like INN...
"You can't always get what you want." [4]
[4] Mick Jagger, German ex-philosopher
Dave Sill
> [Out] of 198 groups, I have no problem with removal for 195 of them.
> That's my opinion. Deal with it.
I appreciate your feedback. Providing my opinion of it is my way of
dealing with it.
-Dave
David Bostwick
>On Wed, 2 Feb 2011 17:40:03 CST, david.b...@chemistry.gatech.edu (David
> Bostwick) wrote in
><04qdnbXCv-6zT9TQ...@earthlink.com>:
>
[...]
>
>>I think, however, that this has become A Mission From God,
>>and will occur no matter what.
>
>I agree that it is almost certainly going to occur.
>
>I guess we'll find out whether it makes any difference
>after it's done. "One observation is worth 10,000
>expert opinions." ;o)
>
> Marty
>
"What should we do, doctor?"
"We're not sure it's the problem, and it may not help, but we'll take out the
patient's appendix and see if it makes a difference. He probably doesn't need
it anyway. If there's no improvement, we'll try something else. Hand me the
scalpel."
Dave Sill
> "Quality rather than quantity" may be a credo, but I'm not convinced
> that it will be possible only by removing some groups. May be it could
> improve the conditions for further actions.
If you remove all the bad groups, all that remains is good groups.
-Dave
David Bostwick
I don't see a real upside, and it's something that can't be undone. I don't
think it will help, because Usenet's downturn is not because of empty
groups cluttering up the namespace. The group I helped start in 1995,
sci.techniques.mass-spec, has been changed to a mailing list mainly because
Tech decided to decommission the news server and let the software license
expire. (They basically did what the board is now considering, except on
an all-or-nothing scale.) All git.* groups are dead, and we now have no
Usenet access from a campus server. Some universities did that several years
ago, and others are considering it.
If you're not sure something will help, not doing it may be the best course.
I don't believe the potential improvement will materialize. But then, it's
not my call. The board wants to do it, so it will be done.
Alexander Bartolich
> [...]
> "What should we do, doctor?"
>
> "We're not sure it's the problem, and it may not help, but we'll take out the
> patient's appendix and see if it makes a difference. He probably doesn't need
> it anyway. If there's no improvement, we'll try something else. Hand me the
> scalpel."
All right, that's a funny way to suggest pathological behavior.
My counter move is less subtle.
http://en.wikipedia.org/wiki/Compulsive_hoarding
# Compulsive hoarding (or pathological hoarding or disposophobia)
# is the excessive acquisition of possessions (and failure to use or
# discard them), even if the items are worthless, hazardous, or un-
# sanitary.
Ciao
Alexander.
Steve Bonine
> "What should we do, doctor?"
>
> "We're not sure it's the problem, and it may not help, but we'll take out the
> patient's appendix and see if it makes a difference. He probably doesn't need
> it anyway. If there's no improvement, we'll try something else. Hand me the
> scalpel."
Very funny. And a perfect analogy, except that we're not putting the
patient under anesthesia.
Seriously, David . . . what's your objection? Do you think that having
a newsgroup list full of dead groups is somehow a good thing? Does
removing them have any downside risk at all? The only objection you've
raised is that you don't think that it's worth doing, but you're not
being asked to do it.
I understand a response that points out a risk related to the project,
or a response that explains why a given newsgroup should be retained.
Objection just for the sake of objecting is a great tradition of Usenet,
and if that's what you're doing I'll file it away in the appropriate
receptacle.
Harald Maedl
Uuuh, if you would remove all the bad in the world, all that remains
would be nothingness. CfV for nothing.at.all now! And none of your
excuses like "all" wouldn't be an allowed part of a group name;)
David Bostwick
>David Bostwick wrote:
>> [...]
>> "What should we do, doctor?"
>>
>> "We're not sure it's the problem, and it may not help, but we'll take out the
>
>> patient's appendix and see if it makes a difference. He probably doesn't
> need
>> it anyway. If there's no improvement, we'll try something else. Hand me the
>
>> scalpel."
>
>All right, that's a funny way to suggest pathological behavior.
>My counter move is less subtle.
>
My intent wasn't to show psychological problems, just to point out that an
irreversible action with no promise of improvement (and possibly no connection
to the problem) may be a bad idea.
>http://en.wikipedia.org/wiki/Compulsive_hoarding
># Compulsive hoarding (or pathological hoarding or disposophobia)
># is the excessive acquisition of possessions (and failure to use or
># discard them), even if the items are worthless, hazardous, or un-
># sanitary.
>
Ah, but empty groups are neither hazardous nor unsanitary, and there have been
some statements that at least some are not worthless.
David Bostwick
I've already replied to the question in a different post. Basically, the
culling is being done because some people think it's A Good Thing, and that
keeping the empty groups is A Bad Thing. I don't automatically share either
of those opinions, just as I didn't agree with every aspect of the Great
Transfer. The decline of Usenet will continue with the empty groups retained
or culled, so maybe I'm just playing Devil's Advocate.
Out of the 7 billion people in the world, those of us discussing this are less
than a blip. I've expressed my opinion and given some reasons, and those in
control of the keys will do what they want. EOT.
David E. Ross
> REQUEST FOR DISCUSSION (RFD)
>
> This is a formal Request for Discussion (RFD) to remove the following
> 198 unmoderated newsgroups.
>
>
> All groups listed below fulfill these conditions:
> - no moderated groups
> - no group names matching *.misc
> - zero on-topic, non-crossposted threads in the past 18 months
> - on-topic questions that received no on-topic answer do not count
>
> DISTRIBUTION:
>
> news.announce.newgroups
> news.groups.proposals
> news.groups
>
> Because of the magnitude of the group list this proposal is not cross-
> posted to target groups. In the course of these proceedings the B8MB
> will post pointers to this announcement to appropriate groups. Readers
> are encouraged to take initiative and spread the message.
>
>
> The procedure shall take at least 8 weeks, with announcements posted
> every 4 weeks: 1st RFD, 2nd RFD, and LCC. The group lists may be re-
> vised during this stage. Discussion about candidate groups should take
> place in moderated group news.groups.proposals. After publication of
> the LCC the board votes on each newsgroup individually.
> More details can be found here:
>
> http://www.big-8.org/wiki/Mass_removal_of_groups
>
> NEWSGROUP LINES:
[snipped]
> sci.aquaria Only scientifically-oriented postings about aquaria.
> sci.engr.coastal Coastal and ocean engineering.
> sci.engr.marine.hydrodynamics Marine Hydrodynamics.
> sci.med.occupational Repetitive Strain Injuries (RSI) & job injury issues.
> soc.adoption.parenting Adoptive parenting by adoptive parents.
> soc.college.teaching-asst Issues affecting collegiate teaching assistants.
> soc.support.depression.seasonal Seasonal affective disorder.
>
> CHANGE HISTORY:
>
> 2011-02-01 1st RFD
>
I have reviewed all of the sci.* and soc.* newsgroups listed in the RFD.
In all cases but one, no on-topic messages were visible through either
Google Groups or Giganews with posting dates after 1 July 2009. The one
exception was sci.med.occupational, which had two on-topic announcements
visible through Google Groups but not through Giganews.
I sent a message to each group I reviewed, cross-posted to not more than
three other relevant groups. The message indicated by name that the
target group was being considered for elimination and referred to this
RFD. Each such message had Followup-To set to this newsgroup
(news.groups.proposals).
--
David E. Ross
<http://www.rossde.com/>
Anything I post in this newsgroup is my personal
opinion and does not reflect the official position
of the Big8-Usenet Board.
Harald Maedl
> Ah, but empty groups are neither hazardous nor unsanitary,
Football teams are playing with only one ballooned ball. Pls picture, if
every player would play with his own flat an empty ball. In earlier
times they have tried this with only one empty head, but even this
proved failure.
Alexander Bartolich
> Alexander Bartolich <alexander...@gmx.at> wrote:
>> [...]
>> All right, that's a funny way to suggest pathological behavior.
>> My counter move is less subtle.
>
> My intent wasn't to show psychological problems, just to point out that an
> irreversible action with no promise of improvement (and possibly no connection
> to the problem) may be a bad idea.
Why do you consider this action irreversible? Servers that act on
rmgroup controls will in general also adhere to newgroups.
I guess with "problem" you mean the decline of users and traffic. Well,
I have never claimed that removal of dead groups will improve on that.
Actually I am quite fatalistic and think that Usenet will inevitably
shrink to BBS levels.
>> http://en.wikipedia.org/wiki/Compulsive_hoarding
>> # Compulsive hoarding (or pathological hoarding or disposophobia)
>> # is the excessive acquisition of possessions (and failure to use or
>> # discard them), even if the items are worthless, hazardous, or un-
>> # sanitary.
>
> Ah, but empty groups are neither hazardous nor unsanitary, and there
> have been some statements that at least some are not worthless.
Go be a messie somewhere else. I want my Usenet to be tight.
Ciao
Alexander.
Alexander Bartolich
> [...]
> I have reviewed all of the sci.* and soc.* newsgroups listed in the RFD.
Thank you.
> [...]
> I sent a message to each group I reviewed, cross-posted to not more than
> three other relevant groups. The message indicated by name that the
> target group was being considered for elimination and referred to this
> RFD. Each such message had Followup-To set to this newsgroup
> (news.groups.proposals).
Wonderful!
Please note that I am trying to keep track of pointer posts:
http://www.big-8.org/wiki/The_Great_Downsizing_2011/1
--
host -t mx moderators.isc.org
Doug Freyburger
> David Bostwick wrote:
>
>
> To me, a managed hierarchy has only one asset: the canonical list of
> newsgroups. The flip side of adding of newsgroups is removing them.
That's the key to me. If the Big-8 are managed heierarchies then they
should have maintenance practiced on them. Think of it as the seasonal
pruning in a managed garden.
To me it's a separate question to ask if it will work. Maybe it will
increase fragmentation but the fact is the groups on the list have seen
no traffic for long enough that fragmentation should not matter. When
you divide zero by an integer the result is still zero.
Maybe it will result in a measurable change in traffic patterns maybe it
will not. But to claim heirarchies are managed without taking action to
maintain them makes the claim meaningless. It's the right thing to do.
Character is what you do when no one is watching - In this case posting.
Mark Kramer
Steve Bonine <s...@pobox.com> wrote:
>I'd like to turn this around. Do you have an objection to removing dead
>groups from the list? If someone else wants to identify newsgroups with
>no activity and prune them from the canonical list, is there a downside
>to that? It might even improve the quality of discussion in the
>remaining ones.
I am still at a loss to understand why you think that removing a group
that has NOBODY reading or posting to it will in any way change anything
about any discussions anywhere else. If you are right about the number of
people involved in a dead group, nobody will notice, and nobody will
have any reason to do anything differently.
Is this a fear kind of thing? Do you think other groups are going to react
to this mass removal by upping their level of traffic "just in case" they
are the next target? Is that TRULY a good thing for Usenet? Or will it
scare them away to the web where they don't have to come up with stuff
to say just so someone won't shut the system off on them?
>So if there's no downside, and potential improvement from doing it, why not?
There is a downside and little potential improvement, that's why.
Mark Kramer
Alexander Bartolich <alexander...@gmx.at> wrote:
>Why do you consider this action irreversible? Servers that act on
>rmgroup controls will in general also adhere to newgroups.
I consider it irreversible because I know that there is more to the
process of getting it back than just "send a newgroup".
Once the group is killed as being a "dead topic", it will be much harder
to get it back through the Big 8 process than if it were an original one.
One of the first responses to a request for a newsgroup on the topic
will be "we removed it a year ago because it was dead, why will now be
any different?" Then it requires hitting the sometimes moving target
of evidence required to justify a group on that target today. (We are
fighting a battle of "too few proposals" already -- is forcing people
to go through that process to get BACK what they HAD really justifiable?)
If the server the new readers use is not automanaged, then they will have
to go beg someone to put it back, and the admin at that point will be
thinking "didn't I just REMOVE that group?"
They will also have to deal with the "a better name would be" discussion,
which might result in it getting a different name than it had in the past,
which puts it all the way back to zero in propogation.
And finally, it is an assumption that rmgroups and newgroups are automated
to the same degree at every server.
>I guess with "problem" you mean the decline of users and traffic. Well,
>I have never claimed that removal of dead groups will improve on that.
That claim has been made here. I'll respond to that claim in a followup
to the person who made it.
>Go be a messie somewhere else. I want my Usenet to be tight.
I like my Usenet the way it is.
Mark Kramer
Rob Kelk <rob...@deadspam.com> wrote:
>On Wed, 2 Feb 2011 14:20:35 CST, c28...@TheWorld.com (Mark Kramer)
>wrote:
>
>>In article <ydpqrb4...@UBEblock.psr.com>,
>>Winston <w...@ubeblock.psr.com.invalid> wrote:
>>> If one only cares about identifying "dead" groups, another approach is
>>>one that doesn't involve end-user statistics at all -- many smaller news
>>>servers only carry groups their users have requested, or that look active.
>>>Start by removing newsgroups they carry from the list of all newsgroups.
>>
>>There is a much simpler method: the user can figure it out for himself,
>>using his own criteria.
>
>Here's a criterion: "If I'm not subscribed to the group, it's a waste
>of space."
Ok. That's your opinion, and I have no problem with you holding it. If
you run a server, feel free not to carry it.
>Since I'm a user of Usenet, if you're accepting the users'
>criteria, when are we getting rid of the 2207 excess Big-8 groups and
>leaving behind the 9 that are useful?
Where did you get the idea that I said that we are leaving up to the
user what groups are being gotten rid of? I said nothing of the kind.
I responded to a statement that began "If one only cares about identifying
'dead' groups,". So, in that context, "leaving it up to the user" is
letting HIM identify what groups HE things are dead, and thus not worth
subscribing to, using his own criteria.
>Leaving this up to the user is not desirable, IMHO.
Leaving this up to the user is both desirable and easier than trying to
come up with Usenet-wide all-inclusive accurate determinations. Why
should your, e.g., criteria be applied globally?
>If we use objective criteria instead of leaving it up to each user, we
>can justify any group removal carried out by pointing to the objective
>criteria and saying the removed groups failed to meet them.
The fact is that we do NOT have to remove a "dead group". There need be
no "objective criteria", we can allow the USER to go to a group, see
nothing of value, and go elsewhere, IF HE WANTS TO. His choice.
>The question here is what those objective criteria should be.
No sir, the first question is why it needs to be done, and only then what
those criteria should be -- if it needs to be done. I say let the user
create his own criteria. Not only does this removal not need to be done,
we do not need "objective criteria" when we don't do it.
>"No
>replies to on-topic posts in 18 months" seems a reasonable criterion to
>me.
And the user is quite capable of determining this for himself. What you
failed to notice is that your "no replies" also includes replies that
were cross-posted. I count that as a reply. I have a different definition
of "no" than others.
>>What this proposal will result in is increased frustration for existing
>>users. Those people who we want to stay here because they will be the
>>anchors for any newbies we can manage to draw here.
>
>Why?
Because they will be scanning the group list (through their newsreader)
every time they come to read news, and they will see that a group they
would read has been removed without notice to them. "Hey, I liked that
topic, I think Usenet ought to have a place for it. Apparently they
don't want me here..."
>Currently, "a new user goes to a new-to-him group and sees NOTHING".
Yep. So it is obvious the group is dead. There is nothing to gain for
him by removing it. There is even an off-chance that asking a question
there that is relevant will result in a reply. Removing that group
removes that chance.
>After the group is removed, the new user won't be able to go
>to that group, and thus won't become frustrated in the first place.
Since there is this major groundswell of existing frustration, please
provide a reference to some quantitative data showing it.
On the other hand, anyone who goes to an empty group can SEE that it
is empty. It's not like all the old posts are hidden, they all show up
(to the length of the expiration on that server) if they are there. If
you post a question in a place that you see nothing else, it is not
unlikely that you will get no answer.
>>How can it not be frustrating for someone who monitors a newsgroup on
>>a regular basis, ready to respond to any discussion that comes up he's
>>interested in, to come by one day and be told by his newsreader that
>>"that group doesn't exist"?
See, that's the answer to your "why".
>If there are no posters in a group, why is anyone lurking there?
Because they haven't unsubscribed. It takes an affirmative action to
unsub. Someone has to actually decide "today I will unsub". I don't do
that, as a general rule, for a group that doesn't show up in my "has
articles" list. It's too much work. But when an article does show up,
I'm there to answer it if I can.
In my case, I have 28 groups in my .newsrc as subscribed. I routinely
see activity in only 4 or fewer groups. It would be more effort than
it was worth to go unsub from the rest, and if something on-topic
shows up there I'll be happy to discuss it. In fact, removing any of
those groups is MORE frustrating for me, because it will trigger an
"this group is invalid, rescanning .newsrc" message that means I will
be waiting for several minutes while the reader discusses the problemm
with the server and finds out how many others are invalid. That's a
waste of my time.
>If
>there are lurkers "ready to respond," why aren't they posting?
Because they have nothing to post? I have nothing to say in 26 of the
groups I subscribe to. Why should I post something there otherwise? (NO,
posting "ping" messages just to keep someone from deciding to remove
the group without warning is NOT a productive use of anyone's time.)
>It doesn't matter what is or isn't included in reader metrics, because
>reader metrics are not an indicator of the health of a group. It
>doesn't matter how many people are ready to read something in a group if
>there's nothing there _to_ read.
But it does counter the argument that someone will be frustrated if they
go to a "dead" group and post something because nobody will read it.
>What's important in the determination of whether a group is healthy is
>poster metrics, not reader metrics.
In your opinion. Readers are part of Usenet, too. In my Usenet, anyway.
Mark Kramer
The "out with the bad air, in with the good" method of CPR was removed
from the curriculum a very very long time ago.
Harald Maedl
> Alexander Bartolich wrote:
>>Why do you consider this action irreversible? Servers that act on
>>rmgroup controls will in general also adhere to newgroups.
> I consider it irreversible because I know that there is more to the
> process of getting it back than just "send a newgroup".
ACK. But before we are thinking about new groups, the B8MB should clean
the house. And then we should think about realizing the vision of the
BIG8 as a quality brand name. And then - probably - new groups could be
useful.
Every warehouse is looking for a place, where is a lot of traffic of
people. Many people are attractive for other people and they will
attract once more people. It's a simple fundamental principle you will
find everywhere in economics. And now the B8MB is trying to capitalize
on this principle.
> Once the group is killed as being a "dead topic", it will be much harder
> to get it back through the Big 8 process than if it were an original one.
ACK, but that is reasonable.
> I like my Usenet the way it is.
Sorry, I'm not. I've never seen, that the monochrome sheet of an empty
newsgroup would have guarenteed satori. And even though it would be a
terrible imagination of thousands of little Usenet Buddhas. There can be
only one!
Rob Kelk
wrote:
>In article <4d49ea7...@news.individual.net>,
>Rob Kelk <rob...@deadspam.com> wrote:
>>On Wed, 2 Feb 2011 14:20:35 CST, c28...@TheWorld.com (Mark Kramer)
>>wrote:
>>
>>>In article <ydpqrb4...@UBEblock.psr.com>,
>>>Winston <w...@ubeblock.psr.com.invalid> wrote:
>>>> If one only cares about identifying "dead" groups, another approach is
>>>>one that doesn't involve end-user statistics at all -- many smaller news
>>>>servers only carry groups their users have requested, or that look active.
>>>>Start by removing newsgroups they carry from the list of all newsgroups.
>>>
>>>There is a much simpler method: the user can figure it out for himself,
>>>using his own criteria.
>>
>>Here's a criterion: "If I'm not subscribed to the group, it's a waste
>>of space."
>
>Ok. That's your opinion, and I have no problem with you holding it. If
>you run a server, feel free not to carry it.
>
>>Since I'm a user of Usenet, if you're accepting the users'
>>criteria, when are we getting rid of the 2207 excess Big-8 groups and
>>leaving behind the 9 that are useful?
>
>Where did you get the idea that I said that we are leaving up to the
>user what groups are being gotten rid of? I said nothing of the kind.
Yes, you did say something of the kind:
>>>There is a much simpler method: the user can figure it out for himself,
>>>using his own criteria.
>I responded to a statement that began "If one only cares about identifying
>'dead' groups,". So, in that context, "leaving it up to the user" is
>letting HIM identify what groups HE things are dead, and thus not worth
>subscribing to, using his own criteria.
>
>>Leaving this up to the user is not desirable, IMHO.
>
>Leaving this up to the user is both desirable and easier than trying to
>come up with Usenet-wide all-inclusive accurate determinations. Why
>should your, e.g., criteria be applied globally?
I never said that it should be. The criterion I offered is the sort of
criteria we get when leaving things up to individual users.
>>If we use objective criteria instead of leaving it up to each user, we
>>can justify any group removal carried out by pointing to the objective
>>criteria and saying the removed groups failed to meet them.
>
>The fact is that we do NOT have to remove a "dead group". There need be
>no "objective criteria", we can allow the USER to go to a group, see
>nothing of value, and go elsewhere, IF HE WANTS TO. His choice.
>
>>The question here is what those objective criteria should be.
>
>No sir, the first question is why it needs to be done,
That question has already been answered by many posters in this thread.
> and only then what
>those criteria should be -- if it needs to be done. I say let the user
>create his own criteria. Not only does this removal not need to be done,
>we do not need "objective criteria" when we don't do it.
>
>>"No
>>replies to on-topic posts in 18 months" seems a reasonable criterion to
>>me.
>
>And the user is quite capable of determining this for himself. What you
>failed to notice is that your "no replies" also includes replies that
>were cross-posted. I count that as a reply. I have a different definition
>of "no" than others.
You are free to run your news server in that fashion.
>>>What this proposal will result in is increased frustration for existing
>>>users. Those people who we want to stay here because they will be the
>>>anchors for any newbies we can manage to draw here.
>>
>>Why?
>
>Because they will be scanning the group list (through their newsreader)
>every time they come to read news, and they will see that a group they
>would read has been removed without notice to them. "Hey, I liked that
>topic, I think Usenet ought to have a place for it. Apparently they
>don't want me here..."
>
>>Currently, "a new user goes to a new-to-him group and sees NOTHING".
>
>Yep. So it is obvious the group is dead. There is nothing to gain for
>him by removing it. There is even an off-chance that asking a question
>there that is relevant will result in a reply. Removing that group
>removes that chance.
>
>>After the group is removed, the new user won't be able to go
>>to that group, and thus won't become frustrated in the first place.
>
>Since there is this major groundswell of existing frustration, please
>provide a reference to some quantitative data showing it.
You want quantitive data for a zero condition???
>On the other hand, anyone who goes to an empty group can SEE that it
>is empty. It's not like all the old posts are hidden, they all show up
>(to the length of the expiration on that server) if they are there. If
>you post a question in a place that you see nothing else, it is not
>unlikely that you will get no answer.
>
>>>How can it not be frustrating for someone who monitors a newsgroup on
>>>a regular basis, ready to respond to any discussion that comes up he's
>>>interested in, to come by one day and be told by his newsreader that
>>>"that group doesn't exist"?
>
>See, that's the answer to your "why".
No, it isn't. My "why" was about "the people we want to stay here", not
about lurkers.
>>If there are no posters in a group, why is anyone lurking there?
>
>Because they haven't unsubscribed. It takes an affirmative action to
>unsub. Someone has to actually decide "today I will unsub". I don't do
>that, as a general rule, for a group that doesn't show up in my "has
>articles" list.
We aren't discussing groups that have articles.
> It's too much work. But when an article does show up,
>I'm there to answer it if I can.
In these cases, articles are not showing up.
>In my case, I have 28 groups in my .newsrc as subscribed. I routinely
>see activity in only 4 or fewer groups. It would be more effort than
>it was worth to go unsub from the rest, and if something on-topic
>shows up there I'll be happy to discuss it. In fact, removing any of
>those groups is MORE frustrating for me, because it will trigger an
>"this group is invalid, rescanning .newsrc" message that means I will
>be waiting for several minutes while the reader discusses the problemm
>with the server and finds out how many others are invalid. That's a
>waste of my time.
And keeping those empty newsgroups is a waste of my disk space. Disks
may be inexpensive, but they aren't free.
>>If
>>there are lurkers "ready to respond," why aren't they posting?
>
>Because they have nothing to post?
If nobody has anything to say in a particular group, why bother keeping
it?
> I have nothing to say in 26 of the
>groups I subscribe to. Why should I post something there otherwise? (NO,
>posting "ping" messages just to keep someone from deciding to remove
>the group without warning is NOT a productive use of anyone's time.)
>
>>It doesn't matter what is or isn't included in reader metrics, because
>>reader metrics are not an indicator of the health of a group. It
>>doesn't matter how many people are ready to read something in a group if
>>there's nothing there _to_ read.
>
>But it does counter the argument that someone will be frustrated if they
>go to a "dead" group and post something because nobody will read it.
If somebody posts to a group, then there's something to read - that
post. That doesn't counter my argument at all.
>>What's important in the determination of whether a group is healthy is
>>poster metrics, not reader metrics.
>
>In your opinion. Readers are part of Usenet, too. In my Usenet, anyway.
What are they reading in empty groups?
Rob Kelk
wrote:
>Not in relation to the specific groups proposed for removal, but as a
>general point...
>
>Doug Freyburger <dfre...@yahoo.com> writes:
>> If you can show that one of the groups on the list has thousands of
>> lurking subscribers I will start to ponder if I should join you in asking
>> to keep it. But I wonder at the value of it. Does it really matter if a
>> group has tens of thousands of silent lurkers?
>
>* It might. Don't forget cases like rec.humor.funny where the lack of
> posts was a moderation problem, not lack of readers. I think it
> entirely possible that that group had tens of thousands of silent
> lurkers. I followed comp.security.announce long after CERT stopped
> posting to it, and even emailed them asking about the lack of posts.
> You, at least, got an answer. I never did.
I think it's telling that the moderation problems in rec.humor.funny
were only seen to be addressed after am unofficial proposal to remove
the group was made.
If it takes a proposal to rmgroup a moderated group - any moderated
group, not just r.h.f - to get the group's moderator team to pay
attention to the group, should that moderation team be involved with
that group?
>* Even with unmoderated groups, I can imagine a situation such as
> alt.fan.la-femme.nikita which had mostly died because the show had ended
> many years ago, but then CW came out with _Nikita_ and the group
> revived. OTOH, newsgroups about ISDN and Amiga OS aren't likely to
> revive once they're dead.
>
>However, as I said, I didn't spot any groups in the list that looked worth
>keeping if they're currently dead.
> -WBE
Kathy Morgan
> Out of the 7 billion people in the world, those of us discussing this are less
> than a blip. I've expressed my opinion and given some reasons,
Thank you for doing that. I'm one of those who feels that lots of empty
groups are a bad thing, so I want to get rid of (most of) the empty
groups. However, I also think that some particular empty (or nearly
empty) groups are of value even if they do appear to be dead, so I
appreciate your comments.
--
Kathy, speaking only for myself
David Bostwick
OK, translate this for me.
David Bostwick
>David Bostwick wrote:
>> Alexander Bartolich <alexander...@gmx.at> wrote:
>>> [...]
>>> All right, that's a funny way to suggest pathological behavior.
>>> My counter move is less subtle.
>>
>> My intent wasn't to show psychological problems, just to point out that an
>> irreversible action with no promise of improvement (and possibly no
> connection
>> to the problem) may be a bad idea.
>
>Why do you consider this action irreversible? Servers that act on
>rmgroup controls will in general also adhere to newgroups.
>
OK, probably irreversible. People aren't willing to jump through hoops any
more. The board will have fewer and fewer things to do, and the Web, or
perhaps mailing lists, will get more of the traffic that used to be here.
>I guess with "problem" you mean the decline of users and traffic. Well,
>I have never claimed that removal of dead groups will improve on that.
A few others have. Culling groups will remove the frustration, and help bring
users into the fold.
>Actually I am quite fatalistic and think that Usenet will inevitably
>shrink to BBS levels.
>
At least we share one opinion. Even the trolls have left.
>>> http://en.wikipedia.org/wiki/Compulsive_hoarding
>>> # Compulsive hoarding (or pathological hoarding or disposophobia)
>>> # is the excessive acquisition of possessions (and failure to use or
>>> # discard them), even if the items are worthless, hazardous, or un-
>>> # sanitary.
>>
>> Ah, but empty groups are neither hazardous nor unsanitary, and there
>> have been some statements that at least some are not worthless.
>
>Go be a messie somewhere else. I want my Usenet to be tight.
>
It's not messy, it's just not the way you want it. Unless I look for a group,
I never see it. As I said, the board wants to cull groups, so it shall be
done.
D Finnigan
> comp.sys.apple2.gno The AppleIIgs GNO multitasking
> environment.
If we may suggest additional groups, I also propose that the following be
removed in addition:
comp.sys.apple2.usergroups
comp.sys.apple2.comm
These groups may have had more merit when they were created nearly 20 years
ago when the Apple II had a larger userbase. Today, they go empty, and the
only postings are the monthly FAQs, which also get posted in the main
comp.sys.apple2 newsgroup. Discussions which would be on-topic in any of
these three groups would be more than welcome (and more widely read) in the
main comp.sys.apple2 newsgroup.
The Apple II community, while still active, has shrunk since these
additional groups were created, and the two newsgroups which are still
active are all that are needed:
comp.sys.apple2 (about 500 posts per month)
comp.sys.apple2.programmer (about 40-60 posts per month)
D Finnigan
> comp.sys.apple2.gno The AppleIIgs GNO multitasking
environment.
I too posted a message to comp.sys.apple2 about this proposal. I have a
feeling that there won't be much opposition to removing csa2.gno. I had
also proposed to remove csa2.comm and csa2.usergroups
John F. Morse
How about the analogy to the Library of Congress, where all the books
are burned that haven't been read in the past X years?
The book burners justify their actions believing nobody will ever read
those books for the knowledge they contain.
Of course Usenet has various retention policies on an individual
decentralized server basis, but some servers do maintain a lengthy
retention.
Are the groups actually dead because no new articles are posted, or are
they really alive because there are lurkers who read to learn and prefer
to keep their mouths shut to not attract trolls and flames?
Or should we rely on Google Groups to retain historical articles, like
Deja News started?
And will centralized Google refrain from deleting/canceling articles
that they feel are offensive, illegal, or disagree with for some reason?
--
John
When a person has -- whether they knew it or not -- already rejected the
Truth, by what means do they discern a lie?
D Finnigan
> and the two newsgroups which are still
> active are all that are needed:
>
> comp.sys.apple2 (about 500 posts per month)
> comp.sys.apple2.programmer (about 40-60 posts per month)
Forgot about comp.sys.apple2.marketplace (about 35 ppm) Sorry. That one is
also a keeper.
(Yes, I realize that it isn't being proposed for removal; I just wanted to
be complete.)
Doug Freyburger
> Harald Maedl <harald...@gmx.de> wrote:
>>David Bostwick wrote:
>>
>>> Ah, but empty groups are neither hazardous nor unsanitary,
>>
>>Football teams are playing with only one ballooned ball. Pls picture, if
>>every player would play with his own flat an empty ball. In earlier
>>times they have tried this with only one empty head, but even this
>>proved failure.
>
> OK, translate this for me.
I took it as a poetic soccer description of dealing with empty or dead
newsgroups. Artistic.
Alexander Bartolich
> [...]
> If we may suggest additional groups, I also propose that the following be
> removed in addition:
>
> comp.sys.apple2.usergroups
> comp.sys.apple2.comm
Yes, indeed, these groups also qualify for the "no on-topic thread in
18 months" rule. However, I would prefer to just take a note for future
removal rounds. The problem of (not) notifying target groups is delicate
enough. I don't want to complicate it further with latecomers.
Also, thank you for the pointer post.
Ciao
Alexander.
Harald Maedl
> are burned that haven't been read in the past X years?
Another analogy: Usenet is like a blackboard. Someone is writing
something on it and one hour later the next one is wiping it out to
write new stories.
> Of course Usenet has various retention policies on an individual
> decentralized server basis, but some servers do maintain a lengthy
> retention.
Do you have examples of those archiving newsservers?
> Or should we rely on Google Groups to retain historical articles, like
> Deja News started?
That's a problem indeed. But you could set up an archive server to store
all news.
Dave Sill
>
> How about the analogy to the Library of Congress, where all the books
> are burned that haven't been read in the past X years?
The difference is that books have content. If the proposal was to to burn
all of the unread empty books in the LOC that would be a better analogy.
But even that's not a perfect analogy.
-Dave
Doug Freyburger
> John F. Morse wrote:
>
>> How about the analogy to the Library of Congress, where all the books
>> are burned that haven't been read in the past X years?
>
> The difference is that books have content. If the proposal was to to burn
> all of the unread empty books in the LOC that would be a better analogy.
> But even that's not a perfect analogy.
A better analogy is going through the Dewey decimal system sorted
cataloging and noticing what topics have had no new books published in
many years. Then closing those numbers in the catalog and no longer
searching for new books on those topics.
In the case of the google deja news archive, closed groups are still
searchable. They can no longer be posted to. No one is frustrated by
lack of response by the mechanism of not being able to post. It's a
proactive response to a closed topic.
Kathy Morgan
> If we may suggest additional groups, I also propose that the following be
> removed in addition:
>
> comp.sys.apple2.usergroups
> comp.sys.apple2.comm
Thank you. I think there will be further removals, so the suggestion is
helpful.
--
Kathy, member of B8MB but speaking only for myself.
Kathy Morgan
> > decentralized server basis, but some servers do maintain a lengthy
> > retention.
>
> Do you have examples of those archiving newsservers?
I believe Giganews has an extremely long retention period in text
groups. Sadly, they reportedly also have many bogus groups and don't do
a particularly good job of filtering spam.
--
Kathy
D Finnigan
> Step 2: Gather stats on injection servers, see which are most popular,
> these are going to be the biggest reading servers.
>
It ends up looking something like this, in descending order:
1.) Google Groups
2.) Eternal-September
3.) Giganews
These are the Big-3 of Usenet today.
D Finnigan
Giganews claims a retention of over 2,783 days for text groups. That works
out to be 7 years ago, back to June of 2003.
No word, however, on whether all text newsgroups have 7 years' retention.
They might have been expiring the really high-traffic groups back then.
A great server for high-retention text groups is textnews.cambrium.nl
This server is read-only, and retention varies depending on group traffic.
Some newsgroups have retention back to 2006. Most go back to at least 2008.
Only the highest-traffic newsgroups have less than a year of articles
stored.
Mark Kramer
Harald Maedl <harald...@gmx.de> wrote:
>Another analogy: Usenet is like a blackboard. Someone is writing
>something on it and one hour later the next one is wiping it out to
>write new stories.
Epic fail. Similar to the epic fail of the "everyone in a room talking
at once" analogy. My posting this article does nothing to "wipe out"
any previous article; my posting this article does not prevent you from
reading ("hearing") any other article.
The existance of a "dead group" does not take volumes of space on critical
and expensive disks, nor does it consume limited bandwidth to transmit. It
does not prevent other groups from having other discussions, nor will
the removal of a dead group have any effect on any other group.
>> Or should we rely on Google Groups to retain historical articles, like
>> Deja News started?
>
>That's a problem indeed. But you could set up an archive server to store
>all news.
I have archives of some newsgroups existing from the day I started
archiving them -- almost two decades ago.
Mark Kramer
Doug Freyburger <dfre...@yahoo.com> wrote:
>In the case of the google deja news archive, closed groups are still
>searchable. They can no longer be posted to. No one is frustrated by
>lack of response by the mechanism of not being able to post.
Really? So you would not be frustrated if you find a newsgroup precisely
on the topic you have a question you desperately need and answer for,
can't find the specific answer in the archive of that group, and then find
out that you cannot ask your question of the people who might have the answer
you need because a group of people who have no interest in that topic and don't
read the group decided to remove it because it was "a closed topic"?
Hmmm. I think I would be frustrated. I might consider Usenet a "closed
topic" at that point and not bother with it further. Are we looking for
new users, or do we want to "close" the topics that some may want to come
here for?
Paul W. Schleck
Hash: SHA1
> REQUEST FOR DISCUSSION (RFD)
>RATIONALE:
>DISTRIBUTION:
>news.announce.newgroups
>news.groups.proposals
>news.groups
>PROCEDURE:
> http://www.big-8.org/wiki/Mass_removal_of_groups
>NEWSGROUP LINES:
[snipped]
>sci.aquaria Only scientifically-oriented postings about aquaria.
>sci.engr.coastal Coastal and ocean engineering.
>sci.engr.marine.hydrodynamics Marine Hydrodynamics.
>sci.med.occupational Repetitive Strain Injuries (RSI) & job injury issues.
>soc.adoption.parenting Adoptive parenting by adoptive parents.
>soc.college.teaching-asst Issues affecting collegiate teaching assistants.
>soc.support.depression.seasonal Seasonal affective disorder.
>CHANGE HISTORY:
>2011-02-01 1st RFD
>--
I have reviewed all of the rec.* newsgroups listed in the RFD. In all
cases but three, no on-topic threads of discussion (versus individual
articles with no replies) were visible through either Google Groups or
Giganews with posting dates after 1 July 2009. The exceptions
(rec.arts.sf.tv.quantum-leap, rec.crafts.dollhouses, and
rec.music.artists.stevie-nicks) are elaborated below.
I sent a message to each group I reviewed, cross-posted to not more than
four other groups. The message indicated by name that the target group
was being considered for elimination and referred to this RFD. Each
such message had Followup-To set to this newsgroup
(news.groups.proposals).
I visited each newsgroup using the nn newsreader, with kill and init
files disabled, using the Newscene news server, which has retention
between 30 and 90 days depending on the newsgroup. This was mostly to
see if there was any very recent activity, whether archived at Google or
not.
I then used Google Groups to look at longer newsgroup histories, usually
back to the start of each newsgroup. There is the risk that I am
missing articles that had a "X-No-Archive: yes" header inserted by
posters, but figure that the probability of this is low, and the general
usage pattern of a newsgroup can be reasonably determined by the likely
much larger sample of archived messages. One advantage of Google Groups
over other alternatives is that they offer a quick tabular view of the
month-by-month posting counts going back to the start of a newsgroup's
lifetime by selecting the "More Group Info" link on a newsgroup's topic
list page. Many newsgroups started out with high activity, but over
time, dropped off to almost nil. Some groups failed early, sometimes
well before Usenet's general decline, some groups petered out around
2005-2006, some around 2007-2008.
Summary descriptions of the activity in each newsgroup are as follows:
rec.arts.comics.elfquest
Only activity since February of 2009 has been Chinese designer clothing
and wristwatch SPAM. Last significant on-topic activity was back in
late 2008. Activity dropped off to a handful of posts a month as far
back as 2005.
rec.arts.sf.starwars.collecting.customizing
Only activity since June of 2009 has been Chinese designer clothing,
shoe, and wristwatch SPAM. Last significant on-topic activity was back
in late 2008. Activity dropped off to a handful of posts a month as far
back as 2002 or 2003.
rec.arts.sf.tv.quantum-leap
Only a recent discussion thread started by KalElFan (Feb 2) in response
to the Downsizing RFD. Only meta-discussion in the thread. The rest of
the activity since about spring of 2008 has been Chinese designer
clothing, shoe, and wristwatch SPAM. Last significant on-topic activity
was back in late 2007 (though there have been some strange "Pic Quiz
Result" articles posted without explanatory context or followup through
spring of 2008, which I am not sure are on-topic or not). Activity
dropped off to a handful of posts a month as far back as 2006.
rec.autos.sport.rally
Nearly all activity since 2008 has been Chinese designer clothing, shoe,
drug, porn, and dietary supplement SPAM. Last significant on-topic
activity was back in early 2008. Activity dropped off to a handful of
posts a month as far back as 2006.
rec.crafts.dollhouses
Only one recently posted two-article discussion thread (Feb 3) started
by "nerobi10", possibly in response to the Downsizing RFD. No reply.
Most recent SPAM was for wristwatches. On-topic traffic was mostly
single article announcements about products or services. Last on-topic
discussion thread was from spring of 2008. Last significant on-topic
activity was from fall of 2007. Activity dropped off significantly past
2006, with most months since then having 1 or no articles.
rec.games.bolo
Nearly all activity since early 2007 has been Chinese designer clothing,
shoe, drug, porn, and dietary supplement SPAM. Last significant
on-topic activity was back in early 2007. Activity dropped off to as
few as a handful of posts a month as far back as 2001 or 2002.
rec.games.computer.quake.editing
Nearly all activity since early 2007 has been SPAM, mostly for
cigarettes, on-line (non-Quake) games, and pirated software, with little
or no designer clothing, shoe, drug, porn, or dietary supplement SPAM.
Last on-topic thread with followups was in spring of 2007. Last
significant on-topic activity was way back in summer of 2004. Activity
dropped off to as few as a handful of posts a month as far back as 2003.
rec.games.computer.quake.playing
Nearly all activity since early 2007 has been SPAM, mostly for on-line
(non-Quake) games, and pirated software, with little or no designer
clothing, shoe, drug, porn, or dietary supplement SPAM. Last on-topic
thread with followups was in fall of 2006. Last significant on-topic
activity was way back in late 2004. Activity dropped off to as few as a
handful of posts a month as far back as 2002 or 2003.
rec.games.computer.quake.quake-c
Nearly all activity since late 2002 has been SPAM, mostly for on-line
(non-Quake) games, financial/banking scams, and pirated software, with
little or no designer clothing, shoe, drug, porn, or dietary supplement
SPAM. Last on-topic thread with followups, as well as any significant
on-topic activity, was in late 2002. Activity dropped off to as few as
a handful of posts a month as far back as 2001 or 2002, with at least
half of the months since 2005 having no posted articles at all.
rec.games.frp.industry
Noticeably less SPAM than the other rec.games.* newsgroups. Last
on-topic thread with followups was in fall of 2009, though it was a
large thread (178 articles). This newsgroup may have a high number of
active lurkers who might post if there is activity. Activity dropped
off to as few as a handful of posts a month as far back as 2003, with at
least half of the months since 2003 having no posted articles at all.
rec.games.frp.storyteller
Only one recent article, from Todd Rich, possibly in response to the
Downsizing RFD. No reply. Noticeably less SPAM than the other
rec.games.* newsgroups. Last on-topic thread with followups, as well as
any significant activity, was in fall of 2008. This newsgroup may have
a high number of active lurkers who might post if there is activity.
Activity dropped off to as few as a handful of posts a month as far back
as 2005, with at least half of the months since 2005 having no posted
articles at all.
rec.games.trading-cards.marketplace.magic.trades
This is a marketplace (swap) newsgroup, so there may not be much
followup discussion by custom or charter. Nearly all of the recent SPAM
is wristwatch-related. The last on-topic for-sale ads for Magic trading
cards appear to be in summer of 2006. Activity dropped off to as few as
a handful of posts a month as far back as 2005, with at least half of
the months since 2008 having no posted articles at all.
rec.games.trading-cards.startrek
Strangely enough, nearly all of the recent SPAM is wristwatch-related.
The last on-topic discussion threads appear to be in fall of 2007.
Activity dropped off to as few as a handful of posts a month as far back
as 2002 or 2003, with at least half of the months since 2008 having no
posted articles at all.
rec.games.video.cd-i
Nearly all activity since 2007 has been SPAM, mostly for modded game
consoles, and on-line games. The last on-topic thread with followups,
as well as any significant on-topic activity, was in summer of 2007.
Activity dropped off to as few as a handful of posts a month as far back
as 2003, with at least half of the months since 2007 having no posted
articles at all. The CD-I format appears to be obsolete.
rec.games.video.cd32
Nearly all activity since 2007 has been SPAM, mostly for wristwatches
and pirated games. Very few single-article on-topic posts since 2001 or
2002, and even fewer discussion threads. Activity on this newsgroup had
not been much more than 1 or 2 posts a day since 1995. Activity dropped
off to as few as a handful of posts a month as far back as 2001, with at
least half of the months since 2005 having no posted articles at all.
The CD32 format appears to be obsolete.
rec.games.xtank.play
Continuous domination of newsgroup by off-topic SPAM and
tangentially-related (non-xtank) computer game announcements since the
mid-1990's. The last on-topic article with followups appears to be from
summer of 2001. A handful of single-article on-topic posts over most of
the newsgroup's lifetime. Activity dropped off to as few as a couple of
posts a day as far back as the start of the newsgroup in 1992, with at
least half of the months since 2005 having no posted articles at all.
The xtank game (presumably part of the X11 software distribution)
appears to be either obsolete or a very obscure sub-interest of
X11-based games (itself probably an obscure sub-interest).
rec.games.xtank.programmer
Less recent SPAM than rec.games.xtank.play. Since this newsgroup has
"programmer" in its name, there was a brief spate of programmer job
announcements around 2000. Last on-topic thread appears to be from fall
of 1999. A handful of single-article on-topic posts over most of the
newsgroup's lifetime. A weekly XTank FAQ was posted by Bill Woddall
through about early 1997. Activity dropped off to as few as a couple of
posts a day as far back as the start of the newsgroup in 1992, with at
least half of the months since 2001 having no posted articles at all.
The xtank game (presumably part of the X11 software distribution)
appears to be either obsolete or a very obscure sub-interest of
X11-based games (itself probably an obscure sub-interest).
rec.music.artists.ani-difranco
Mostly wristwatch SPAM since early 2008. The last on-topic threads with
any followup discussion appear to be from fall of 2005. The last
significant on-topic activity appears to be from 2004. Activity dropped
off significantly much after mid 2003. About half of all months since
2007 had 1 or no articles.
rec.music.artists.danny-elfman
Less SPAM than other newsgroups. A handful of on-topic, single posts
over most of the newsgroup's lifetime. The last significant on-topic
activity appears to be from 2002. Activity dropped off significantly
much after mid 1999. About half of all months since 2004 had 1 or no
articles
rec.music.artists.stevie-nicks
Most recent SPAM, going back to about 2008, has been for jewelry and
wristwatches. The last significant on-topic activity appears to be from
summer of 2009. Very low activity much past 2003, mostly related to
occasional touring dates and news about Fleetwood Mac.
rec.music.artists.wallflowers
Less SPAM than other newsgroups, mostly wristwatches. Last on-topic
post with a followup was from early 2009. Last significant on-topic
activity was from fall of 2007. Very low activity much past 2001,
mostly related to occasional touring dates and news about related bands.
rec.music.iranian
Much less SPAM than other newsgroups. Continuous, but low, on-topic
activity up until about summer of 2009. Mostly single-post
announcements about Iranian/Persian music. Nearly all multi-post
discussion threads were flame-wars about Iran's government and politics.
Low activity over the newsgroup's lifetime, tapering off significantly
much after 2001, and with most months after 2008 having 1 or no
articles.
rec.parks.theme
Low incidence of SPAM, both recently, and over the newsgroup's lifetime.
Last on-topic discussion thread was from summer of 2008. Mostly single
article announcements and trip reports through 2007. Low activity
newsgroup whose traffic dropped off significantly much after 2007, with
many months having 1 or no articles.
rec.scouting.guide+girl
Some recent SPAM over the last few years, but not a high amount or
general pattern. Last on-topic discussion thread was from summer of
2007. Mostly single articles, often relayed from news media or
organizational press releases through 2007. Low activity newsgroup
whose traffic dropped off significantly much after 2006, with many
months having 1 or no articles.
rec.sport.basketball.women
Some recent Indian porn and wristwatch SPAM over the last few
years. Last on-topic discussion thread was from spring of 2009. Some
single-article for-sale posts relating to memorabilia through 2009.
More single article announcements than discussion threads over the most
recent few years. Activity dropped off significantly much after 2004,
with many months having 1 or no articles.
rec.sport.footbag
Last article, which was SPAM, was from late 2009. Mostly wristwatch
SPAM. Likely not any on-topic discussion threads since 2003 or 2004,
with only a small amount of on-topic single articles after that. Low
activity newsgroup whose traffic dropped off much after 2001, with many
months having 1 or no articles.
rec.sport.orienteering
Recent traffic was mostly designer clothing and shoe SPAM. Last
on-topic discussion thread was from summer of 2008. Last significant
on-topic activity was from fall of 2007. Low-to-medium activity
newsgroup whose traffic dropped off significantly much past 2001, with
many months having 1 or no articles.
rec.sport.skating.roller
Recent traffic was mostly designer clothing, shoe, and wristwatch SPAM.
Last on-topic discussion thread was from late 2008. Last significant
on-topic activity was from mid-to-late 2007. Low traffic newsgroup
whose activity gradually tapered off since 2000, dropping significantly
past 2005, with many months having 1 or no articles.
rec.video.dvd.advocacy
Recent traffic was mostly wristwatch and pirated DVD/game SPAM. Last
on-topic discussion thread was from spring of 2005. Very little
significant on-topic activity over the newsgroup's lifetime, mostly used
as a crossposting target for general DVD information and assistance.
Very little activity much past 2001, with many months after 2006 having
1 or no articles.
- --
Paul W. Schleck
psch...@novia.net
http://www.novia.net/~pschleck/
Finger psch...@novia.net for PGP Public Key
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (SunOS)
iD8DBQFNTL0a6Pj0az779o4RAntEAJ9rFqgVgE2/DknD6dPWffOdzF9q4wCfUIOE
1Imqkl2WdsNI3OnYmtuyJR4=
=/3kl
-----END PGP SIGNATURE-----
Steve Bonine
> Doug Freyburger wrote:
You are making the implicit assumption that if you posted your question
in this "newsgroup precisely on the topic" that you would receive an
answer. Available data suggests the opposite; many of the newsgroups
proposed for removal have questions that were posted to them for which
no one responded. The frustration level will be higher if the
individual takes the time to compose a submission to the newsgroup and
no answer is forthcoming.
Unless the individual is on Google and knows exactly the name of this
"newsgroup precisely on the topic", they're not going to see it. Much
more likely is the scenario in which they choose a newsgroup that's
close to the topic of the question. It is more likely that they will
receive an answer to a question posted there than in a perfectly named
but unfortunately dead newsgroup.
Thank you for providing a concrete illustration of how removal of one
newsgroup affects other newsgroups.
Brad Templeton
>>
>> To me, a managed hierarchy has only one asset: the canonical list of
>> newsgroups. The flip side of adding of newsgroups is removing them.
>
>That's the key to me. If the Big-8 are managed heierarchies then they
>should have maintenance practiced on them. Think of it as the seasonal
>pruning in a managed garden.
>
>To me it's a separate question to ask if it will work. Maybe it will
>increase fragmentation but the fact is the groups on the list have seen
>no traffic for long enough that fragmentation should not matter. When
>you divide zero by an integer the result is still zero.
>
>Maybe it will result in a measurable change in traffic patterns maybe it
>will not. But to claim heirarchies are managed without taking action to
>maintain them makes the claim meaningless. It's the right thing to do.
>Character is what you do when no one is watching - In this case posting.
It reminds me of the debates, never settled on the idea of "no votes"
in the grand newsgroup voting system. Many people asked "why are
there no votes?" Why did it matter if people didn't like a group,
so long as enough people wanted it?
The usual answer was that no votes allowed people to not like a name,
or occasionally a policy, or a need to censor a dangerous topic
but I must say I never found them satisfying.
Designing a taxonomy by participatory democracy is a clear FAIL in
my view.
The truth is that the "100 more yes than no" rule was pretty much
made up on a whim, and the whole voting system was primarily there
to convince those who wanted a pointless newsgroup that few people
really supported them. The NO votes may have been another way to
do that.
The issue here is not large, but the issue of what powers it gives
to a board are larger than the value or lack-of-value of a set of
dead or dormant groups. I suspect the debate is not so much over
the value of the groups, but about whether this step is of sufficient
value to establish the precedent of mass rmgroup.
I would not be too bothere by it myself, if it weren't for the fact
that I entered this group because those involved were preparing to
modify my own group under the thesis that they could not find me, in
spite of my being, to be frank, about as findable as can be on the net.
It did not inspire confidence in the process as constituted, I am
afraid.
--
Sail up the Yangtse river into China in my photojournals
http://www.templetons.com/brad/photo/china/