woods@ncar.ucar.edu (Greg Woods) (03/24/89)
I have been informed by Spaf that he no longer maintains the official
guidelines for creation of newsgroups. I would like to take over that duty
(Why me? Because I'm relatively well-known to the net, I'm willing to put
in the time required and deal with the inevitable flames that will result, and
I volunteered first :-)
First a plea: this is probably the most important topic we could possibly
discuss in news.groups. I am cross-posting it to news.admin simply because
the news administrators are the ones who will have to "enforce" the
guidelines so they should be in on this discussion. The plea is, because
this topic is so important, let's for once try to concentrate on the issues
at hand and avoid personal attacks. I welcome honest criticism of these
proposed guidelines. To do my part in keeping down the flames, I will not
respond to any article that looks more like a personal attack on me than
a criticism of the guidelines, no matter how angry I get or how untrue
the statements of a personal nature might be. I encourage everyone else
to do the same.
Second, a disclaimer: This is only a DRAFT, folks. It isn't carved in
stone, nor do I expect it to be accepted without some modifications.
What this is is a formal statement of the rules that I see actually being
followed on the net. It is true that some of these rules are the subject of
controversy. I am trying my best to reflect the net consensus, but, since I
am a human being and therefore have personal opinions just like everyone else,
I cannot guarantee that this statement of the guidelines is 100% free of my
personal biases. In fact, where no clear consensus has been reached it is
quite likely that my personal opinions entered into the formal statement.
At any rate, here it is. I welcome private or posted comments and will
be posting a new updated version that will reflect the consensus of those
comments (as best I can) in a week or so.
RULES FOR USENET GROUP CREATION AND DELETION (draft)
REQUIREMENTS FOR GROUP CREATION:
These guidelines should be followed when you think a new newsgroup should
be created. Any group creation request which follows these guidelines to
a successful result should be honored, and any request which fails to follow
these procedures or to obtain a successful result from doing so should be
dropped, except under extraordinary circumstances. The reason these are
called guidelines and not absolute rules is that it is not possible to
predict in advance what "extraordinary circumstances" are or how they might
arise.
The Discussion
1) A call for discussion on creation of a new newsgroup should be posted
to news.groups, and also to any other groups or mailing lists at all
related to the proposed topic if desired. The Followup-to: header should be
set so that the actual discussion takes place only in news.groups
2) The discussion period should last for at least two weeks (14 days).
3) The name of the proposed group must be agreed upon during the discussion
period, which may last up to 30 days if needed.
The Vote
1) AFTER the discussion period, if it is determined that a new group is
really desired and a name and charter are agreed upon, a call for
votes may be posted to news.groups and any other groups or mailing
lists that the original call for discussion might have been posted
to. The call for votes should include clear instructions for how to
cast a vote. It must be as clearly explained and as easy to do to
cast a vote for creation as against it, and vice versa. It is
explicitly permitted to set up two separate addresses to mail yes
and no votes to provided that they are on the same machine, to set
up an address different than that the article was posted from to
mail votes to, or to just accept replies to the call for votes
article, as long as it is clearly and explicitly stated in the call
for votes article how to cast a vote.
2) The voting period must last for exactly 30 days, no matter what the
preliminary results of the vote are.
3) A couple of repeats of the call for votes may be posted during the vote,
provided that they contain similar clear, unbiased instructions for
casting a vote as the original, and provided that it is really a repeat
of the call for votes on the SAME proposal (see #5 below).
4) ONLY votes MAILED to the vote-taker will count. Votes posted to the net
for any reason (including inability to get mail to the vote-taker) and
proxy votes (such as having a mailing list maintainer claim a vote for
each member of the list) are NOT ALLOWED.
5) Votes may not be transferred to other, similar proposals. A vote shall
count only for the EXACT proposal that it is a response to. In particular,
a vote for or against a newsgroup under one name shall NOT be counted as
a vote for or against a newsgroup with a different name or charter.
6) Votes MUST be explicit; they should be of the form "I vote for the
group foo.bar as proposed" or "I vote against the group foo.bar
as proposed". The wording doesn't have to be exact, it just needs to
be unambiguous. In particular, statements of the form "I would vote
for this group if..." should be considered comments only and not
counted as votes.
The Result
1) At the completion of the 30 day voting period, the vote taker must post
the vote tally and the names and/or E-mail addresses of the votes received
to news.groups and any other groups or mailing lists to which the original
call for votes was posted.
2) AFTER the vote result is posted, there will be a 3 day waiting period
during which the net will have a chance to correct any errors in
the voter list or the voting procedure.
3) AFTER the waiting period, if 100 more YES/create votes were received
than NO/don't create, a newgroup control message may be sent out.
If the 100 vote margin is not met, the group should not be created and
discussion of it on the net should cease.
RULES FOR DELETING A USENET GROUP
[Since there has obviously been no consensus ever reached on this, this
statement cannot be construed as anything other than my personal opinion]
I propose that the procedure be the same as for group creation, with
100 more keep than delete votes required to save the group, with the exception
that all calls for votes and discussions MUST be cross-posted to both
news.groups and the group in question. This will obviously only work for
unmoderated groups; I am open to discussion on procedures for deleting
moderated groups, but I would prefer to wait for a final resolution on
the rec.humor.funny/Brad Templeton controversy before trying to formalize
them, because it will be IMPOSSIBLE to keep the flames down until then.david@ms.uky.edu (David Herron -- One of the vertebrae) (03/24/89)
I have one change to suggest --
That of the total number of votes, the yes votes need be greater than
some percentage (66%? 75%?). (While still being >100 vote margin)
This is for votes which have high numbers of no votes but still a
large enough number of yes votes. When the numbers are large
then the relative percentages get skewed. If it's
yes: 400
no: 299
then the vote would pass even though the margin is pretty close.
As the numbers get larger these margins get closer and closer.
And in case anybody thinks that such high numbers of votes aren't
possible, all you need is one volatile subject (like comp.women)
to set it off and you'll get lots of votes.
My feeling is that in such a case, even though a lot of people have
indicated favourably, that a *LOT* of people have *ALSO* indicated
unfavorably. And that group of people is large enough that they
should be given attention. That there must be something fundamentally
wrong with the proposal if it's generating that much negative feelings.
*REGARDLESS* of the amount of positive feelings.
And please, let's not re-hash the comp.??.women argument. Yes this
idea did come up in that context, but it's not apropos to my suggestion
here.
--
<- David Herron; an MMDF guy <david@ms.uky.edu>
<- ska: David le casse\*' {rutgers,uunet}!ukma!david, david@UKMA.BITNET
<-
<- The problem with mnemonics is they mean different things to different people.chuq@Apple.COM (Chuq Von Rospach) (03/24/89)
>That of the total number of votes, the yes votes need be greater than >some percentage (66%? 75%?). (While still being >100 vote margin) Good idea, but let's keep it simple. A given vote must have 50% + 1 of the vote being for creation, with a minimum number of votes (say, 250). That shows that (1) the majority of the net wants it created, and (2) that there's a relatively broad-base of support. Those are more vigorous requirements than it takes to be elected to government office! Chuq Von Rospach -*- Editor,OtherRealms -*- Member SFWA chuq@apple.com -*- CI$: 73317,635 -*- Delphi: CHUQ -*- Applelink: CHUQ [This is myself speaking. No company can control my thoughts.] USENET: N. A self-replicating phage engineered by the phone company to cause computers to spend large amounts of their owners budget on modem charges.
jbuck@epimass.EPI.COM (Joe Buck) (03/24/89)
In article <11326@s.ms.uky.edu> david@ms.uky.edu (David Herron -- One of the vertebrae) writes: >I have one change to suggest -- > >That of the total number of votes, the yes votes need be greater than >some percentage (66%? 75%?). (While still being >100 vote margin) While your proposal has some merit, the version that Greg Woods posted has the advantage of representing the current consensus. The three day waiting period hasn't been specified explicitly before, but it's basically the way things worked, since we generally waited until Spaf, Greg, or some other generally recognized "important news honcho" posted the newsgroup. In short, we already have a workable consensus on how to create groups, and Greg has just spelled out existing practice. Cases such as David describes -- 400 vs 299 -- have never happened. When we get large numbers of votes it gets extremely one-sided. I know they are possible, but remember that since every news administrator is sovreign over his/her machines, the rules only work if everyone assents to them. How long would it take before everyone agrees to some new formula? Let's leave newsgroup creation rules as they are, since we already have a rule that basically works. (Oh, by the way, I applaud the choice of Greg Woods to fill Spaf's shoes). However, we've never had a newsgroup deletion rule before, and I have one major concern about Greg's suggestion. Since a majority of 100 "yes" over "no" votes is required to KEEP a group, I'm concerned about the possibility of "vote harrassment" -- essentially frivolous votes on smaller groups where the people that benefit from that group have to keep voting again and again. (Bob Webber, are you still out there? Think what fun you could have with this rule!). I'll cogitate on this one for a while and see if I can come up with a better idea. -- -- Joe Buck jbuck@epimass.epi.com, uunet!epimass.epi.com!jbuck
sob@harvisr.harvard.edu (Scott Bradner) (03/24/89)
>6) Votes MUST be explicit; they should be of the form "I vote for the > group foo.bar as proposed" or "I vote against the group foo.bar > as proposed". The wording doesn't have to be exact, it just needs to > be unambiguous. In particular, statements of the form "I would vote > for this group if..." should be considered comments only and not > counted as votes. I don't like this one. It is much easer to follow the results if a format like: YES - fred@bill.foobar.edu NO - fred@joe@foobar.edu were permitted, I do agree that the ambigueous should be discarded. Scott
csu@alembic.UUCP (Dave Mack) (03/24/89)
In article <1634@ncar.ucar.edu> woods@ncar.UCAR.EDU (Greg Woods) writes: [Rules for newsgroup creation - well-stated, no particular surprises.] > > RULES FOR DELETING A USENET GROUP > >[Since there has obviously been no consensus ever reached on this, this > statement cannot be construed as anything other than my personal opinion] This should come as no surprise then. >I propose that the procedure be the same as for group creation, with >100 more keep than delete votes required to save the group, with the exception >that all calls for votes and discussions MUST be cross-posted to both >news.groups and the group in question. I think that demonstrated disinterest, indicated by "inadequate" volume, should also be a component of this formula. Otherwise, it will be possible to dismantle groups with high volume and a large readership on the basis of a vote. This strikes me as a bad, bad idea. It would lead to newgroup- rmgroup wars that would make anything in the past look like a tea party. I suggest that a reasonable basis for newsgroup removal would be that the group has had fewer than some threshhold number of articles per month (5-10?) for the preceding three months, as measured at three "hub" sites (say rutgers, uunet and decwrl?) The hub sites must, of course, carry the group in question over the measurement period. This requirement would have to be met before the removal vote could be taken. > This will obviously only work for >unmoderated groups; I am open to discussion on procedures for deleting >moderated groups, You need to deal with three options separately: 1) forcibly switching moderators 2) changing the group from moderated to unmoderated over the objection of the current moderator 3) removing the group altogether And the discussion must be carried on where those most affected will be sure to see it, in the moderated group. Obviously, this presents a problem for many posters. One solution would be to have the Keeper of the List (you, if you become the new Keeper) post a notice of the discussion with an Approved: line in the moderated group, notifying the readers of the nature of the discussion and advising the readers to tune into news.groups for a while. I look forward to the next installment in this saga. -- Dave Mack cannibal dancer
nyssa@terminus.UUCP (The Prime Minister) (03/24/89)
In article <1634@ncar.ucar.edu> woods@ncar.UCAR.EDU (Greg Woods) writes: >3) AFTER the waiting period, if 100 more YES/create votes were received > than NO/don't create, a newgroup control message may be sent out. > If the 100 vote margin is not met, the group should not be created and > discussion of it on the net should cease. While the rest is good, I think the time has come to examine this 100 vote plurality rule. Since the net has been growing, and still is growing, perhaps a more generic formula could be made, relating the needed plurality to size of the net?
cik@l.cc.purdue.edu (Herman Rubin) (03/24/89)
In article <27773@apple.Apple.COM>, chuq@Apple.COM (Chuq Von Rospach) writes: > > Good idea, but let's keep it simple. A given vote must have 50% + 1 of the > vote being for creation, with a minimum number of votes (say, 250). That > shows that (1) the majority of the net wants it created, and (2) that > there's a relatively broad-base of support. Those are more vigorous > requirements than it takes to be elected to government office! So you would support creating a newsgroup if 126 were for creation and 124 against, but not if 126 were for and 0 against? I fail to see the logic of this. I do not see 126 for and 124 against as better than 110 for and 10 against, which would pass the current requirement. Thus, I see Chuq proposing a procedure which looks worse than what we are now doing. -- Herman Rubin, Dept. of Statistics, Purdue Univ., West Lafayette IN47907 Phone: (317)494-6054 hrubin@l.cc.purdue.edu (Internet, bitnet, UUCP)
win@gatech.edu (Win Strickland Jr) (03/24/89)
In article <1634@ncar.ucar.edu> woods@ncar.UCAR.EDU (Greg Woods) writes: > > > RULES FOR USENET GROUP CREATION AND DELETION (draft) > >REQUIREMENTS FOR GROUP CREATION: I like the way Greg has stated these requirements. Minor variations to his proposal would be ok, but in general, I think they capture what we are doing now. > RULES FOR DELETING A USENET GROUP > >[Since there has obviously been no consensus ever reached on this, this > statement cannot be construed as anything other than my personal opinion] > Here, I agree with some comments made by others, and I think we should have some sort of traffic measure which causes this process to initiate rather than some person deciding it would be fun to see if a vote would fail. I think Dave Mack proposed some sort of volume requirements over a three month period. Again, we might want to discuss the specific numbers in terms of month and volume (5, 10, ? articles), but in general I think this might be the way to go. I've always wanted to see some sort of guidelines for deleting groups, and I hope we can reach a consensus on this. -- Win Strickland Jr School of ICS, Georgia Tech, Atlanta GA 30332 Internet: win@gatech.edu root@gatech.edu UUCP: ...!{decvax,hplabs,ncar,purdue,rutgers}!gatech!win Phone: 404 894 3086 (msgs 404 894 3152)
compton@silver.bacs.indiana.edu (Dave Compton) (03/24/89)
In article <11326@s.ms.uky.edu> david@ms.uky.edu (David Herron -- One of the vertebrae) writes: :I have one change to suggest -- : :That of the total number of votes, the yes votes need be greater than :some percentage (66%? 75%?). (While still being >100 vote margin) : :This is for votes which have high numbers of no votes but still a :large enough number of yes votes. When the numbers are large :then the relative percentages get skewed. If it's : : yes: 400 : no: 299 I have another suggestion. Either a vote > 100 vote margin, or 250 or more yes votes. Say there is a call for votes for a newsgroup with a controversial subject. And for the sake of argument, the vote come out to. Yes 300 No 299 Now, In my humble opinion, this newsgroup has as much right to exist as a speciality group that passes with the vote of: Yes 105 No 4 After all, the creation of a newsgroup is not as important as electing a president. So if enough people want the group, let them have it even if there is not a 2/3 majority or >100 yes to no votes. dave -- ------------------------------------------------------------------------------- | compton@silver.bacs.indiana.edu | These are only my opinions, | | rutgers!iuvax!silver!compton@iuvax | but I'm an opinionated person! | -------------------------------------------------------------------------------
roy@phri.UUCP (Roy Smith) (03/25/89)
I pretty much agree with Greg Woods's proposed rules. I only have
one comment; I think the "100 more Yes than No" rule is not strong enough.
I don't remember exactly when the 100 threshold was thought of, but I'm sure
the net was 1/Nth the size it is today, where N is in the interval (2,5).
Perhaps the threshold could float, being defined as X% of the people the
latest arbitron reports as being on the net. I leave it to the netnews
lawyers to hash out the details of how we determine X. My off-the-cuff
guess is that something on the order of 1% is probably reasonable.
--
Roy Smith, System Administrator
Public Health Research Institute
{allegra,philabs,cmcl2,rutgers,hombre}!phri!roy -or- roy@phri.nyu.edu
"The connector is the network"Ram-Ashwin@cs.yale.edu (Ashwin Ram) (03/25/89)
In article <3088@alembic.UUCP>, csu@alembic.UUCP (Dave Mack) writes: > In article <1634@ncar.ucar.edu> woods@ncar.UCAR.EDU (Greg Woods) writes: > > > > RULES FOR DELETING A USENET GROUP > > > >I propose that the procedure be the same as for group creation, with > >100 more keep than delete votes required to save the group, with the exception > >that all calls for votes and discussions MUST be cross-posted to both > >news.groups and the group in question. > > I think that demonstrated disinterest, indicated by "inadequate" volume, > should also be a component of this formula. [...] > I suggest that a reasonable basis for newsgroup removal would be that > the group has had fewer than some threshhold number of articles per > month (5-10?) for the preceding three months [...] I'm not sure I like this. There are a few serious low-volume newsgroups that it would be a shame to delete purely on the basis of "inadequate" volume (e.g., some of the bionet newsgroups; the t-discussion list would fall into this category if it were gatewayed to a newsgroup; etc.) These newsgroups are useful, and their low volume makes them all the more manageable. (Now for talk or soc newsgroups the volume requirement makes sense :-)). I'm also not convinced that "100 more keep than delete votes to save the group" is a good idea, since the default here seems to be that the group would be deleted unless its readers band together to save it. If this is to be the default, shouldn't there be some requirement for initiating a call to delete a group in the first place? It seems unreasonable that anyone could "threaten" an existing group merely by posting a call for its deletion, forcing its readers to start sending in "keep" votes. Having said this, I agree that there should be a way to judge lack of interest in a particular newsgroup, based on volume and readership among other things. However, either of these criteria by itself doesn't seem to be sufficient by itself. -- Ashwin.
dlm@cuuxb.ATT.COM (Dennis L. Mumaugh) (03/25/89)
In article <1634@ncar.ucar.edu> woods@ncar.UCAR.EDU (Greg Woods) writes: I have been informed by Spaf that he no longer maintains the official guidelines for creation of newsgroups. I would like to take over that duty (Why me? Because I'm relatively well-known to the net, I'm willing to put in the time required and deal with the inevitable flames that will result, and I volunteered first :-) ... These guidelines should be followed when you think a new newsgroup should be created. Any group creation request which follows these guidelines to a successful result should be honored, and any request which fails to follow these procedures or to obtain a successful result from doing so should be dropped, except under extraordinary circumstances. The reason these are called guidelines and not absolute rules is that it is not possible to predict in advance what "extraordinary circumstances" are or how they might arise. Exceptional conditions are those that are caused by events such that a reasonable person (Aunt Martha) would predict will cause enough net traffic and discusssion to require a private news group: for example, comp.sys.next, or sci.space.shuttle. Sometimes the major need is to focus discussions in various other news groups and/or agree on the name. 1) A call for discussion on creation of a new newsgroup should be posted to news.groups, and also to any other groups or mailing lists at all related to the proposed topic if desired. The Followup-to: header should be set so that the actual discussion takes place only in news.groups The call for discussion should show that there is a need for the group by citing traffic in other groups that would justify it. Unless there is already traffic one shouldn't propose a new group just for the sake of completeness of the name space. 2) The discussion period should last for at least two weeks (14 days). Or longer as net propogation and weekends tend to make it possible for a person to not see a call for discussion. Also not all people are news addicts and some people have vactions. 3) The name of the proposed group must be agreed upon during the discussion period, which may last up to 30 days if needed. And if it isn't? 2) The voting period must last for exactly 30 days, no matter what the preliminary results of the vote are. I object to exactly 30 days. At least 30 days and not more than 37 days. Allow for email propagation, bounced mailers and mail hops that go down because of facilities failure (att-ih went down for a couple of days recently and I got no feed until it was fixed.). Also allow for news being expired and a person not knowing until the last of the voting period. 4) ONLY votes MAILED to the vote-taker will count. Votes posted to the net for any reason (including inability to get mail to the vote-taker) and proxy votes (such as having a mailing list maintainer claim a vote for each member of the list) are NOT ALLOWED. Can we have more than one vote taker (with the various takers cooperating). Could somebody like uunet offer as a net service a automatic vote taking service for each vote? 1) At the completion of the 30 day voting period, the vote taker must post the vote tally and the names and/or E-mail addresses of the votes received to news.groups and any other groups or mailing lists to which the original call for votes was posted. Posting of results should be done on a Monday (see below). 2) AFTER the vote result is posted, there will be a 3 day waiting period during which the net will have a chance to correct any errors in the voter list or the voting procedure. See above about propagation and email problems. Suggest 5 days or 3 working days with announcements on a Monday. 3) AFTER the waiting period, if 100 more YES/create votes were received than NO/don't create, a newgroup control message may be sent out. If the 100 vote margin is not met, the group should not be created and discussion of it on the net should cease. Suggest that the results be announced on a Monday and the group creation be on the next Monday. This can be reduced or eliminated iff the vote is lopsided -- e.g. 400 to 3. RULES FOR DELETING A USENET GROUP [Since there has obviously been no consensus ever reached on this, this statement cannot be construed as anything other than my personal opinion] I propose that the procedure be the same as for group creation, with 100 more keep than delete votes required to save the group, with the exception that all calls for votes and discussions MUST be cross-posted to both news.groups and the group in question. This will obviously only work for unmoderated groups; I am open to discussion on procedures for deleting moderated groups, but I would prefer to wait for a final resolution on the rec.humor.funny/Brad Templeton controversy before trying to formalize them, because it will be IMPOSSIBLE to keep the flames down until then. I recommend different. A group is subject to deletion if it has been inactive for more than 3 months (unless it is deemed a "permanent group" as is news.announce.all). Inactive is determined by the following "central" locations not seeing traffic: decwrl, uunet, and rutgers. Or, any other diverse sites. The group is proposed as a candidate for deletion in news.groups and the group itself. A cross post to the next highest .misc group (if such exists) is also done. [Followup to news.groups only]. There is another wait for 30 days before a vote is taken. If articles other than chatter on why the group exists or other such meta-discusions appear on the group in question the proposal should be aborted, OR the group be diverted into the next higher level .misc group via aliases. Then the vote to be taken must have a minimum of 100 votes for retention and need not obtain any votes for execution. It is executed iff there aren't enough votes to sustain its quiet non-existence. Traffic in the group is sufficient to stop the vote [I'll beg to question on people making traffic to justify the group]. I suggest adding a new section on the mechanics of group creation or deletion: 1). When a new group is created, the newgroup control message will contain a copy of the (corrected) vote. One day after the creation, an article will be posted with the welcome message and a (re)statement of the groups charter. 2). A rmgroup control message will contain a justification for the removal and an enumeration of the steps to verify the need for removal. -- =Dennis L. Mumaugh Lisle, IL ...!{att,lll-crg,attunix}!cuuxb!dlm OR dlm@cuuxb.att.com
fred@cdin-1.UUCP (Fred Rump) (03/25/89)
The creation part sounds like it's always been and seems to be reasonably
enforceable. No further comment there.
But on the deletion of newsgroups there should be some statistical tool that
initiates the actual deletion process. As we all know it is easy to approve
but much more difficult to reverse the process. We need a VALID reason that
reasonable netters can respect and be somewhat logical about.
If, for instance, group ABC has not had enough activity over a pre-determined
timeframe, the deletion process should automatically commence. The number of
votes to delete should NOT be as stringent as the number to create as many
people won't bother voting to delete something they have no interest in. On
the other hand the fever of the discussion to create may have instigated
enough interest to vote YES for the benefit of those who clamored for it. But
now if those who demanded a new group have not sufficiently used it, let's
correct a hasty mistake quickly and painlessly.
The alternative is an evergrowing set of special interest groups that may
become a slow end of the net. We do have mail where a handful of interested
individuals can communicate with each other without broadcasting to the world.
Perhaps they were able to meet each other thru a newsgroup which then fizzled
past the point of continuing interest for all but a handful of individuals. So
let them continue, but not in news.
Perhaps some of the more knowledgeable counters of postings can shed some
light as to how many messages or posters they consider to be an active group.
Somehow I think the number of active posters should be a critical factor in
the deletion discussion. If the 'readers' never respond, maybe there isn't
much to say?
Fred Rump
...uunet!cdin-1!fred
--
Fred Rump, Pres. | UUCP: {rutgers,cbmvax,bellcore}!bpa!cdin-1!fred
CompuData, Inc. | or ...{allegra killer gatech!uflorida decvax!ucf-cs}
10501 Drummond Rd. | !ki4pv!cdis-1!cdin-1!fred
Philadelphia, Pa. 19154| or ...!uunet!cdin-1!fred or fred@cdin-1.uu.netfwebb@bbn.com (Fred Webb) (03/25/89)
Greg, if you're willing to take on this thankless job, more power to you! In general, I think your proposed guidelines are fine, and are pretty close to what's been happening. I do have a couple of minor comments. 1. On your proposed waiting period - I'm not sure 3 days is anywhere near enough, given net propagation delays. I would say 10 days would be the minimum to ensure that all objections have been seen. Of course, that may seem like too long, especially if the vote was overwhelming. I don't know what to do here. Anybody have an idea? 2. On deletion - I think it should be harder to delete a group than it was to create it in the first place. On the other hand, if nobody cares enough to vote, the default should probably be to delete. I propose an alternate set of guidelines: 50 votes to keep are enough to keep the group, and votes to delete aren't counted at all. So, if a substantial number of people want to keep the group, it can't be taken away from them. The rest of what you said about deletion seems just about right. 3. It seems to me that deleting a moderated group should follow the same guidelines. What we need is some way to ensure that the call for votes gets out, to the group, even over the moderator's objection. 4. Finally, it seems to me that we may need a way to force replacement of the moderator of a moderated group. If the moderator stops moderating, and messages to him go unanswered, I think it should just happen. Maybe you, as keeper of the "Official Guidelines" could decide that this situation exists, issue a call for volunteers for moderator, and call for votes on the new moderator. Again, messages would have to go to the group about this. Replacing a moderator who doesn't want to be replaced probably just shouldn't happen. -- Fred
aem@ibiza.Miami.Edu (a.e.mossberg) (03/25/89)
In <11326@s.ms.uky.edu>, <david@ms.uky.edu> wrote: >I have one change to suggest -- >That of the total number of votes, the yes votes need be greater than >some percentage (66%? 75%?). (While still being >100 vote margin) That sounds like a good suggestion. Perhaps: YesVotes > 100 && YesVotes > 3*NoVotes aem a.e.mossberg aem@mthvax.miami.edu MIAVAX::AEM (Span) aem@umiami.BITNET (soon) Beardsley's Warning to Lawyers: Beware of and eschew pompous prolixity.
chuq@Apple.COM (Chuq Von Rospach) (03/25/89)
>1. On your proposed waiting period - I'm not sure 3 days is anywhere near >enough, given net propagation delays. I would say 10 days would be the minimum >to ensure that all objections have been seen. I'd leave it at some small, reasonable time. Three days isn't unreasonable. If you look at the propogation times from uunet, most messages get through the backbone very quickly, and it's very unlikely that a site far enough out not to see a message in three days is going to be a site that's in the middle of a controversy. Remember, if a group is created and a controversy erupts, it can always be un-created if it was found to have been created incorrectly. There is precedent for that. There's no reason to wait unless there's a general feeling that the vote is going to be challenged, and I think it's a safe bet we'll know in advance what votes are going to be controversial. >2. On deletion - I think it should be harder to delete a group than it was to >create it in the first place. On the other hand, if nobody cares enough to >vote, the default should probably be to delete. I propose an alternate set of >guidelines: 50 votes to keep are enough to keep the group, and votes to delete >aren't counted at all. Make the size of the required margin dependent on the volume. A group that is rarely/never used should be easier to get rid of than a group that has a high volume. And I disagree that it should just be "keep" votes. Put the burden of proof on the people who want to delete a group by making the default to keep it. Therefore, the vote should be some percentage of votes. Hmm. A thought just occurred to me. The "100 more yes than no votes" has always been somewhat controversial because the absolute number is relatively easy to overcome. Thinking about it, a better way is to define succeed/fail in terms of percentages. How about this: o A group creation vote succeeds if 67% of the vote is for creation. o A group deletion vote succeeds if 67% of the vote is for deletion. o In both cases, if the number of spoiled or proven missing ballots exceeds 2% of the total vote, or if enough spoiled/missing votes are found that would have changed the result of the vote, the vote must be re-taken by an agreed-upon neutral third party. o Any vote with less than 150 total votes will be considered failed for lack of interest. This puts the bias against creating groups and against deleting existing ones. A group that cares about it's survival should be able to roust enough votes to stay above the 33% limit. And it sets up the success to be relative to the total interest in the group rather than trying to set some arbitrary number (with a low-end limit set in to keep ennui from taking hold). >3. It seems to me that deleting a moderated group should follow the same >guidelines. What we need is some way to ensure that the call for votes gets >out, to the group, even over the moderator's objection. >4. Finally, it seems to me that we may need a way to force replacement of the >moderator of a moderated group. Whenver there is a controversy over a moderated group or a moderator, I think it's critical that any voting be monitored by a neutral third party. That way neither side can claim bias and muddle the proceedings with their noisy whimperings. Also, there should be procedures for turning groups moderated and unmoderated, as well as changing moderators.
welty@steinmetz.ge.com (richard welty) (03/25/89)
i think that by and large, these are reasonable guidelines. i do,
however, agree with dave mack and some others that the proposed
deletion guidelines make a very unpleasant sort of harrassment
possible.
also, it would be nice if persons voting on a proposal were
required to provide something resembling a real name in their
votes -- i always feel a little queasy counting votes from
`VHDL Project' or `superuser' or `guest' or whatever;
i'd like some evidence of a person behind the particular vote.
this still won't protect votes from rabid sysadministrators
sending in votes from newly made-up accounts, but there isn't
anything i can think of that would protect against that form
of abuse.
richard
--
richard welty 518-387-6346, GE R&D, K1-5C39, Niskayuna, New York
welty@steinmetz.ge.com welty@crd.ge.com uunet!steinmetz!welty
``It is General Electric policy to maintain neutrality on the question
as to whether P = NP'' -- Bob Mattheysesoconnor@nuke.steinmetz (Dennis M. O'Connor) (03/25/89)
An article by welty@steinmetz.ge.com (richard welty) says: ] it would be nice if persons voting on a proposal were ] required to provide something resembling a real name in their ] votes ... i'd like some evidence of a person behind the particular vote. ] this still won't protect votes from rabid sysadministrators ] sending in votes from newly made-up accounts ... Nor from systems like Lisp machines, which often have no security at all usually ( although it IS possible to activate it ). All you need to create an account on a Symbolics Lispm is attempt to login. These insecure machines are on the net, in fact, Dick Welty used one to post his article. Well, maybe it was Dick Welty : our Lispms don't usually have passwords on accounts enabled. It could have been anyone. -- Dennis O'Connor oconnor%sungod@steinmetz.UUCP ARPA: OCONNORDM@ge-crd.arpa "Jethro Tull won the Heavy Metal/Hard Rock Grammy. Clearly, I am a MetalHead."
mack@inco.UUCP (Dave Mack) (03/25/89)
In article <54708@yale-celray.yale.UUCP> Ram-Ashwin@cs.yale.edu (Ashwin Ram) writes: >In article <3088@alembic.UUCP>, csu@alembic.UUCP (Dave Mack) writes: >> In article <1634@ncar.ucar.edu> woods@ncar.UCAR.EDU (Greg Woods) writes: >> > >> > RULES FOR DELETING A USENET GROUP >> > >> >I propose that the procedure be the same as for group creation, with >> >100 more keep than delete votes required to save the group, with the exception >> >that all calls for votes and discussions MUST be cross-posted to both >> >news.groups and the group in question. >> >> I think that demonstrated disinterest, indicated by "inadequate" volume, >> should also be a component of this formula. [...] >> I suggest that a reasonable basis for newsgroup removal would be that >> the group has had fewer than some threshhold number of articles per >> month (5-10?) for the preceding three months [...] > >I'm not sure I like this. There are a few serious low-volume newsgroups that it >would be a shame to delete purely on the basis of "inadequate" volume (e.g., >some of the bionet newsgroups; the t-discussion list would fall into this >category if it were gatewayed to a newsgroup; etc.) These newsgroups are >useful, and their low volume makes them all the more manageable. (Now for talk >or soc newsgroups the volume requirement makes sense :-)). You deleted the last sentence in that paragraph. I said that the low volume requirement had to be met before you could vote on removal, not that low volume resulted in removal. >I'm also not convinced that "100 more keep than delete votes to save the group" >is a good idea, since the default here seems to be that the group would be >deleted unless its readers band together to save it. If this is to be the >default, shouldn't there be some requirement for initiating a call to delete a >group in the first place? It seems unreasonable that anyone could "threaten" an >existing group merely by posting a call for its deletion, forcing its readers to >start sending in "keep" votes. Agreed. It should be 100 more delete than keep votes, not the other way around, with at least 100 delete votes necessary. Or whatever. >Having said this, I agree that there should be a way to judge lack of interest >in a particular newsgroup, based on volume and readership among other things. >However, either of these criteria by itself doesn't seem to be sufficient by >itself. Perhaps that is the function of the vote? Basically, we would be asking for a vote of confidence in the group. -- Dave Mack
cc1@valhalla.cs.ucla.edu (It glows in the dark) (03/25/89)
In article <3520@silver.bacs.indiana.edu> compton@silver.UUCP (Dave Compton) writes:
^
^ I have another suggestion. Either a vote > 100 vote margin,
^or 250 or more yes votes. Say there is a call for votes for a
^newsgroup with a controversial subject. And for the sake of argument,
^the vote come out to.
^
^Yes 300
^No 299
^
^ Now, In my humble opinion, this newsgroup has as much right to
^exist as a speciality group that passes with the vote of:
^
^Yes 105
^No 4
^
This is just ridiculous. 300 to 299 vote passes? What about 299
to 300? What about 299 to 3,000? That's more than 250 yes votes.
This has to be the stupidest posting yet that I've seen on this
subject.
A little thought, please, before giving out your ideas to the net.
--Kenbrendan@jolnet.ORPK.IL.US (Brendan Kehoe) (03/25/89)
In article <1634@ncar.ucar.edu> woods@ncar.UCAR.EDU (Greg Woods) writes: >4) ONLY votes MAILED to the vote-taker will count. Votes posted to the net > for any reason (including inability to get mail to the vote-taker) and > proxy votes (such as having a mailing list maintainer claim a vote for > each member of the list) are NOT ALLOWED. Hmmmm...in that case, how will we handle getting a vote to the vote-taker if it is impossible for the person to reach that site? (It happens)..asking someone else to vote for you is out..God could people have fun with that.. perhaps we could 1) set up 2 different addresses, the 2nd of which would route mail to the 1st (if mail to #1 is unreliable); or, 2) require that the intended vote-reception site is VERY accessable (now define very). Also, a recommendation: while we're at it, define a standard form of vote reporting (I've seen name/ address, name, sorted by address/name/date of receipt/you_name_it). That way it will be a bit easier to read the reports since we'll have some idea in advance of what il look like. -- Brendan Kehoe brendan@cup.portal.com | GEnie: B.KEHOE | Oh no! I forgot to say goodbye brendan@chinet.chi.il.us | CI$: 71750,2501 | to my mind! brendan@jolnet.orpk.il.us | Galaxy: Brendan | - Abby Normal
dig@peritek.UUCP (Dave Gotwisner) (03/25/89)
In article <3088@alembic.UUCP>, csu@alembic.UUCP (Dave Mack) writes: > In article <1634@ncar.ucar.edu> woods@ncar.UCAR.EDU (Greg Woods) writes: > [Rules for newsgroup creation - well-stated, no particular surprises.] > > > > RULES FOR DELETING A USENET GROUP > > > > I suggest that a reasonable basis for newsgroup removal would be that > the group has had fewer than some threshhold number of articles per > month (5-10?) for the preceding three months, as measured at three > "hub" sites (say rutgers, uunet and decwrl?) The hub sites must, of > course, carry the group in question over the measurement period. This > requirement would have to be met before the removal vote could be taken. > I read this as automated group removal or having a program determine when a group should be removed, and then start the discussion off. The problems that I have with this are the following: 1) What's to say a group might be idle for several month's and then become active again (there are some low volume groups right now that have nothing posted (or at least reaching my site) for one or two months, and then all of a sudden have 10-15 large items posted over the course of the next month, and then cycle back to 0 for one or two months). Add to this, the group volume may be a lot higher than the posting seems. In some groups, very little is posted, but lots of people (may) read the articles. 2) The net traffic for groups without much action would probably increase greatly around times for group removal... Namely, posting an article to start a discussion about removal, the people who read the group (see 1 above) posting articles to say no, prior to the vote, discussions based upon the original discussion but not having anything to do with the original call for discussion (for example, the whole set of discussions currently going on about R.H.F, which started as a discussion about a compilation copyright which was originally derived from the Brad/JDR interactions, etc). Also, add to this the net traffic (ie., mail, not news) when the actual vote gets called. Do this for every one of the groups which either someone or some program puts into a questionable volume catagory, and I can see a VAST increase in net volume (and not just short term, because as the number of groups on the net goes up, the number of groups which have low traffic may also go up). 3) It would only take one person to foul things up and force groups to remain (and skew the list of which groups are used and which aren't). A person could, have a program which might send out a small group of postings (3-4 a week) to his favorite low-volume group just to keep it around. Further, he could do this from several accounts on several different machines (given networks within companies), and noone would know that they weren't for real, without reading them. I recently had to remake inews on my system when the number of active groups was greater than the number that I originally made inews for. Other than that, the disk space occupied by empty directories is not great, nor are the number of inodes, etc. Groups which have nothing posted to them cause no real serious problems to occur (as far as I can tell) (other than eating up a little disk space and one inode per quiet group), that a backup-filesystem / remake-file-system (with larger ilist) / restore-filesystem or that a rmdir/mkdir wouldn't fix. The news that would be generated around removing a group would more than likely eat up more disk than having the empty group around (not to mention the eventual discussion to remake the same group again). I would suggest that we leave group removal up to the site administrators to determine which groups they don't want rather than do a blanket for the net as a whole. If space/inodes are a problem for some systems, they could manage the removal of the group on their system, rather than try something on a net-wide basis which may cause more problems than it would solve. -- ------------------------------------------------------------------------------ Dave Gotwisner UUCP: ...!unisoft!peritek!dig Peritek Corporation ...!vsi1!peritek!dig 5550 Redwood Road Oakland, CA 94619 Phone: 1-415-531-6500
cphoenix@csli.STANFORD.EDU (Chris Phoenix) (03/25/89)
In article <3726@phri.UUCP> roy@phri.UUCP (Roy Smith) writes: > > I pretty much agree with Greg Woods's proposed rules. I only have >one comment; I think the "100 more Yes than No" rule is not strong enough. >I don't remember exactly when the 100 threshold was thought of, but I'm sure >the net was 1/Nth the size it is today, where N is in the interval (2,5). >Perhaps the threshold could float, being defined as X% of the people the >latest arbitron reports as being on the net. I leave it to the netnews >lawyers to hash out the details of how we determine X. My off-the-cuff >guess is that something on the order of 1% is probably reasonable. I don't think this is a good idea. As a user, I don't see most of the net, and I don't care about most of the stuff posted, so why should the size of the net affect whether or not a group gets created? A group now is the same as a group several years ago, no? If a group with 100 members was worthwhile back then, why should it change now? As more people start reading the net, the percentage of readers required to form a group will get smaller, and so there will be more groups. On the other hand, this hasn't killed us yet. If groups are required to have more readers, then they will have a higher volume, and some topics will be squeezed together. From what I've seen, high volume is much more of a problem for users than fragmented topics. (I've been forced to drop several groups already.) So the choice is between letting the net have more groups (annoyance for the administrators) and forcing the groups to gain more readers and posters (major problems for the users). IMHO, the users are more important. 1/2 :-) >-- >Roy Smith, System Administrator >Public Health Research Institute >{allegra,philabs,cmcl2,rutgers,hombre}!phri!roy -or- roy@phri.nyu.edu >"The connector is the network" Chris Phoenix cphoenix@csli.stanford.edu
cphoenix@csli.STANFORD.EDU (Chris Phoenix) (03/25/89)
In article <27806@apple.Apple.COM> chuq@Apple.COM (Chuq Von Rospach) writes: >Make the size of the required margin dependent on the volume. A group that >is rarely/never used should be easier to get rid of than a group that has a >high volume. And I disagree that it should just be "keep" votes. Put the >burden of proof on the people who want to delete a group by making the >default to keep it. Therefore, the vote should be some percentage of votes. Wait a minute. Do we EVER want to delete a group that has a high volume? The only reason I can think of for doing this is if it badly violated net policy, in which case it shouldn't be a votable issue. I agree strongly that the default should be to keep it. I can't see any normal situation in which we would want to delete a group with more than, say, 50 readers and 4 postings/month. These numbers can be derived from arbitron, can't they? As I see it, deletion of a newsgroup should just be a cleanup task. A useful newsgroup should never be deleted (except for violations). As a user, I define 'useful' to be one that people are using. By the time a group should be deleted, the only person who would care enough to call for a vote would be an administrator who wasn't reading it. This seems to be a bad way to start off. >Hmm. A thought just occurred to me. The "100 more yes than no votes" has >always been somewhat controversial because the absolute number is relatively >easy to overcome. Thinking about it, a better way is to define succeed/fail >in terms of percentages. How about this: > >o A group creation vote succeeds if 67% of the vote is for creation. > >o A group deletion vote succeeds if 67% of the vote is for deletion. If it wasn't clear above, I'm opposed to voting on group deletion. >o In both cases, if the number of spoiled or proven missing ballots exceeds > 2% of the total vote, or if enough spoiled/missing votes are found that > would have changed the result of the vote, the vote must be re-taken > by an agreed-upon neutral third party. Given the vagaries of e-mail, this seems a bit harsh. Far more than 2% of my off-system letters get bounced. I'd be very surprised if many votes would pass this test. >o Any vote with less than 150 total votes will be considered failed for > lack of interest. > >This puts the bias against creating groups and against deleting existing >ones. A group that cares about it's survival should be able to roust enough >votes to stay above the 33% limit. One thing that I haven't seen yet--who would vote for deleting a newsgroup? Obviously, most of the people reading it wouldn't. People with a vendetta would, but I would hate to have a process that would be swayed by such a group. The only other people I can think of are administrators. As you are suggesting it, you would basically have to have 100-150 administrators vote against a group to get it deleted. This could quickly stir up resentment among them, for having to go through this every time they want a group deleted, and against them, for trying so hard to delete a group they don't really care about. As far as I can see, this would be forcing lots of people in positions of power to either vote as an unthinking block, or take sides on something they don't care that much about, whenever they want to get something done. For some reason, I just don't like the idea. >And it sets up the success to be relative >to the total interest in the group rather than trying to set some arbitrary >number (with a low-end limit set in to keep ennui from taking hold). IMHO, there are many better ways of determing interest than a vote. [suggestions for moderated groups deleted] I don't know enough about how moderated groups work to have any firm basis for opinion. A few ideas, though: 1) The only reason for deleting a group, moderated or unmoderated, should be total lack of interest or bad violations of rules. In neither case is a vote appropriate. 2) Making a group unmoderated, or changing moderators, is a tough one. Seems like the decision should be up to the group alone. Is there any way of telling who has been subscribed longer than X weeks, and accepting only their votes on such an issue? Chris Phoenix cphoenix@csli.stanford.edu
frank@rsoft.UUCP (Frank I. Reiter) (03/26/89)
In article <27806@apple.Apple.COM> chuq@Apple.COM (Chuq Von Rospach) writes: >o Any vote with less than 150 total votes will be considered failed for > lack of interest. I don't think that a lack of NO votes should in any way contibute to the failure of a vote. I like your "Lack of interest" idea however and suggest that it be based strictly on the number of YES votes. -- _____________________________________________________________________________ Frank I. Reiter UUCP: {uunet,ubc-cs}!van-bc!rsoft!frank Reiter Software Inc. frank@rsoft.bc.ca, a2@mindlink.UUCP Langley, British Columbia BBS: Mind Link @ (604)533-2312, login as Guest
Geva_Apple-Maniac_Patz@cup.portal.com (03/26/89)
As deleting a newsgroup is a drastic step, may I recommend that the burden of the deletion be on those who want it to go (how's THAT for lgalese?) In other words, 100 more 'delete' than 'keep' votes should be required. An exception should be made for sittuations where less than 10 'keep' votes are registered, as this obviously signifies a group with very little interest. Again, this could all be modified to percentages, which I think is a jolly good idea. %%%% Geva %%%% Geva Patz
wbt@cbnews.ATT.COM (William B. Thacker) (03/26/89)
In article <27806@apple.Apple.COM> chuq@Apple.COM (Chuq Von Rospach) writes: > >Hmm. A thought just occurred to me. The "100 more yes than no votes" has >always been somewhat controversial because the absolute number is relatively >easy to overcome. Thinking about it, a better way is to define succeed/fail >in terms of percentages. How about this: > >o A group creation vote succeeds if 67% of the vote is for creation. > >o A group deletion vote succeeds if 67% of the vote is for deletion. > >o Any vote with less than 150 total votes will be considered failed for > lack of interest. I don't like percentages; there are too many loopholes. Consider trying to remove a low-volume group: if the vote goes: Yes=120, No=29 ("Yes" means kill the group) the group stays; less than 150 votes. If, however, you get one more "No" vote, the group is removed. If you think the number of voters is low, you might be better off not voting. I've brought this up before, and gotten a rousing "yawn" in response. Why vote against a group ? I can only think of a few reasons: 1) You think the naming of the group is inappropriate: e.g., sci.humor 2) The number of groups in existence threatens to exceed the capacity of your site's news software. (this is apparently a problem to many non-UN*X sites) 3) The group, by its very nature, is illegal or encourages illegal activity (e.g., rec.terrorism, talk.burglary) such that sites carrying it could incur liability. 4) The group is redundant, the topic being already covered by another group whose volume is not excessive. (e.g., news.announce.important.aprilfirst) IMHO, a lack of established demand or volume is insufficient for a "no" vote; after all, if there's no interest, the group won't get 100 votes. I would hazard that a sizeable fraction, and in many instances, the majority, of "no" votes are cast for reasons I consider inappropriate: 1) "I don't think this topic should be discussed on the net" 2) "There's no point in discussing that" 3) "I consider the topic to be offensive" 4) Racist/sexist/political motivations (i.e., an attempt to repress views one is opposed to) It is, of course, impossible to filter "good" no votes from "bad" ones. I think that, for group retention, all that should be required is some minimum monthly volume; perhaps 15 posting per month, maybe less (remember that some moderated groups are digested, so there's only one posting per day, and none on weekends). I further suggest that , for an existing group, the only legitimate reason to vote "no" is lack of namespace; the other three "valid" reasons I enumerated above should have been addressed when the group was created. If lack of namespace is not a real problem for the bulk of the net, perhaps "no" votes could be disallowed. For group creation, I would suggest that the current guidelines, 100 more "Yes" than "No" votes, are appropriate, but could perhaps be amended to include "in any case, if the total number of "Yes" votes exceeds 200, the vote passes." That is, if enough people want the group as proposed, it should be created, regardless of negative votes. This would further allow a "fast" vote, useful for timely topics; for example, comp.sys.next, as I recall, received over 300 "yes" votes in one week; as soon as 200 "yeses" were received, the vote-taker could call for creation. ------------------------------ valuable coupon ------------------------------- Bill Thacker att!cbnews!wbt "C" combines the power of assembly language with the flexibility of assembly language. Disclaimer: Farg 'em if they can't take a joke ! ------------------------------- clip and save --------------------------------
charlie@mica.stat.washington.edu (Charlie Geyer) (03/26/89)
In article <5134@cbnews.ATT.COM> wbt@cbnews.ATT.COM (William B. Thacker) writes: > I further suggest that , for an existing > group, the only legitimate reason to vote "no" is lack of namespace; > the other three "valid" reasons I enumerated above should have been > addressed when the group was created. If lack of namespace is not > a real problem for the bulk of the net, perhaps "no" votes could be > disallowed. I couldn't agree more. If enough people want a group. Let them have it. Otherwise they will just annoy everyone else by posting in inappropriate groups. The only question is how many is enough? Even lack of namespace is no reason to vote against a group. If lack of namespace is a problem there should be some absolute maximum on the number of groups and as low volume groups get removed proposed new groups have to compete for the available space. But, is this really a problem that can't be solved by sites just not carrying some groups?
chuq@Apple.COM (Chuq Von Rospach) (03/26/89)
>I don't like percentages; there are too many loopholes. percentages are admittedly not perferct. My claim is imply that they are better than pure number boundaries. Perfection is impossible. I'm simply suggesting something that I feel is better. >Consider >trying to remove a low-volume group: if the vote goes: > >Yes=120, No=29 ("Yes" means kill the group) > >the group stays; less than 150 votes. If, however, you get one more >"No" vote, the group is removed. This is a boundary condition. Rules invariably break down at boundary conditions. I could plug any numbers into any voting formula and you could come up with a boundary condition that makes it look bad. So by definition, any set of rules is bad because there's a perturbance at the boundary condition. The idea is to compare the rule against reality and not argue the theories of the edge. In practice, It's very unlikely (I can't think of a case) where the voting is that low *and* that close. Looking back on the last year or so of newsgroup votes, I couldn't find a case that would have been screwed up by the boundary conditions I defined. Therefore, I don't see a problem with the boundary, since in practice it'll never hurt anyone. A second analysis of your case is this: if you can only get 29 votes to save a group, then by definition it doesn't have enough votes to be created (if this were a creation vote instead of a deletion vote) and therefore what you're really showing is a lack of support for the group. The voting has already been biased fairly heavily towards the no-delete side -- you're simply arguing that it should be biased even further, and I don't buy that. >I've brought this up before, and gotten a rousing "yawn" in response. >Why vote against a group ? I can only think of a few reasons: There is another reason, which you evidently haven't heard my lecture on over the last eternity here on the net: 5) You do not want an overly complex namespace. The reason you don't want an overly complex namespace is because it breeds confusion in the naive user. There are two problems with leaving moribund newsgroups around: o A naive user will tend to post to that newsgroup assuming there's an audience. Since the group is moribund, there is by definition a very small audience, causing the user to get a non-existant or minimal response. You end up with either an unhappy user or a posting that gets sent out multiple times trying to find an audience, negatively impacting net volume. o The larger the namespace, the harder it is to figure out what the appropriate newsgroup to post is. In many cases, it's ambiguous what groups a posting belongs. In an efficient namespace, there would be a single appropriate newsgroup for every posting, and cross-posting would be unneccessary. USENET is anything but efficient, and perfect efficiency is an unattainable goal anyway, but we should tweak the namespace towards efficiency where reasonable. This includes both creating newsgroups to handle subjects that aren't covered *and* deleting newsgroups that no longer have a purpose. Think of a moribund newsgroup as USENET's appendix. Generally, you don't see it, you don't think about it, you don't notice it. Every so often, however, something goes wrong with it and then you have problems. With USENET, it is fairly trivial to remove the appendix before it does go wrong, leaving you with a healthier organism in the end. If it took major surgery, I wouldn't recommend this action -- but all it takes is an agreement when it's time to go. If only I'd had that option with my appendix.... Chuq Von Rospach -*- Editor,OtherRealms -*- Member SFWA chuq@apple.com -*- CI$: 73317,635 -*- Delphi: CHUQ -*- Applelink: CHUQ [This is myself speaking. No company can control my thoughts.] USENET: N. A self-replicating phage engineered by the phone company to cause computers to spend large amounts of their owners budget on modem charges.
mack@inco.UUCP (Dave Mack) (03/27/89)
In article <574@peritek.UUCP> dig@peritek.UUCP (Dave Gotwisner) writes: >In article <3088@alembic.UUCP>, csu@alembic.UUCP (Dave Mack) writes: >> In article <1634@ncar.ucar.edu> woods@ncar.UCAR.EDU (Greg Woods) writes: >> [Rules for newsgroup creation - well-stated, no particular surprises.] >> > >> > RULES FOR DELETING A USENET GROUP >> > >> >> I suggest that a reasonable basis for newsgroup removal would be that >> the group has had fewer than some threshhold number of articles per >> month (5-10?) for the preceding three months, as measured at three >> "hub" sites (say rutgers, uunet and decwrl?) The hub sites must, of >> course, carry the group in question over the measurement period. This >> requirement would have to be met before the removal vote could be taken. >> > >I read this as automated group removal or having a program determine when >a group should be removed, and then start the discussion off. No. Read the last sentence of what I wrote. The low volume requirement is a prerequisite for *taking a vote on removal.* Beyond that, you would also need to have removal proposed by someone, a discussion period, and a vote. >I would suggest that we leave group removal up to the site administrators >to determine which groups they don't want rather than do a blanket for >the net as a whole. If space/inodes are a problem for some systems, they >could manage the removal of the group on their system, rather than try >something on a net-wide basis which may cause more problems than it would >solve. I would prefer this solution too, but if we are going to have some guidelines for globally rmgrouping, we should try to reach an agreement on what those guidelines are. Guidelines that half the net disagree with are probably worse than none at all. What we are really talking about are guidelines for removal of a group from the canonical list of "real" newsgroups. No one can force a news admin to remove a group at his site if he doesn't want to. There are valid reasons for trimming down the official newsgroups list, even if it's only to keep the size of a checkgroups message reasonable. -- Dave Mack
shields@yunccn.UUCP (Paul Shields) (03/27/89)
In article <18230@gatech.edu>, win@gatech.edu (Win Strickland Jr) writes: > Here, I agree with some comments made by others, and I think we should > have some sort of traffic measure which causes this process to initiate > rather than some person deciding it would be fun to see if a vote would > fail. I think Dave Mack proposed some sort of volume requirements over > a three month period. Again, we might want to discuss the specific numbers > in terms of month and volume (5, 10, ? articles), but in general I think > this might be the way to go. I think there is a need for deleting some groups even if there is traffic, if only that the name of the group has grown to be obsolete, and some other division is apparent, or that there is mass-confusion in general about the group. (I know of so many users who are infinitely confused about the purpose of some groups, finding that discussions in those groups don't fit the stated purpose, mainly because no one seems to understand what the group is for: What is comp.unix.wizards? What is rec.humor.d? What is the difference between news.admin and news.sysadmin? What is news.misc? Why is there so much junk in misc.all? Are the following language-oriented groups structured correctly? (this list made from grepping for lang in /usr/lib/news/newsgroups.) comp.ai.nlang-know-rep, comp.lang.all, comp.std.c, comp.sys.transputer, sci.lang, sci.lang.japan Who should post where? Why _do_ we have users cross-posting to 7 or more different groups? Does traffic in some groups always consist of crossposted articles? Is this a good thing? ) Some newsgroups are nothing but senseless talk.barbarian diatribe. It seems to me that many of those groups are discussing the SAME things. Why not consolodate them? > I've always wanted to see some sort of guidelines for deleting groups, > and I hope we can reach a consensus on this. Concensus? It'll never happen. But let's see some policy that can at least give us some flexibility. -- Paul Shields, shields@yunccn.UUCP
jha@lfcs.ed.ac.uk (Jamie Andrews) (03/27/89)
Death to boundary conditions! Add 60 "silent no votes" and then require a 60% majority. (Your numbers may vary.) Yes votes ------------------------- >= .60 Yes votes + No votes + 60 ^^^ ^^ Your favourite numbers here Thus if you get 0 no votes, you need 90 yes votes 50 no votes, you need 165 yes votes 100 no votes, you need 240 yes votes 200 no votes, you need 330 yes votes etc., to win. (Knew that frp experience would come in handy! :-)) --Jamie. jha@lfcs.ed.ac.uk "And your flame has burned so very very brightly, Roy"
jj) (03/28/89)
I don't agree at all with the proposed newsgroup deletion rules. I'd suggest the following rules: 1) Newsgroups with no traffic at any major gateway for three months can bite the dust with an announcement to the group, and a two week wait for complaints, given complaints, hold a vote as in 2). 2) Call for a vote, have the two week obligatory discussion period, and THEN remove the group, after a 30 day vote, if 100 more people vote NO than YES, rather than the other way around. Any group may be voted on only ONCE per year for removal. I think any easier sort of voting procedure will lead to "newsgroup wars" the likes of which we've never seen before. I further vote NO for Greg's holding the official list. I've seen his voice strongly stated on one side or the other of too many debates to accept him. The keeper of the list must be a neutral, or close to one. I'm also opposed, but in a lesser way, to the strictures Greg would have us put on group proposals. They seem to me to be nothing but a set of excuses not to create a given newsgroup. -- It's a Small World, After All! *Mail to jj@alice.att.com or alice!jj We're Not Satisfied, At All! *HASA, Athiest Curmudgeon Division Gonna Build One Not So Small *Copyright alice!jj 1989, all rights reserved, except As This Small, Small, World! *transmission by USENET and like free facilities granted.
wbt@cbnews.ATT.COM (William B. Thacker) (03/28/89)
In article <27870@apple.Apple.COM> chuq@Apple.COM (Chuq Von Rospach) writes: >percentages are admittedly not perferct. My claim is imply that >they are better than pure number boundaries. Perfection is impossible. >I'm simply suggesting something that I feel is better. Understood. >>Consider >>trying to remove a low-volume group: if the vote goes: >> >>Yes=120, No=29 ("Yes" means kill the group) >> >>the group stays; less than 150 votes. If, however, you get one more >>"No" vote, the group is removed. > >This is a boundary condition. Rules invariably break down at boundary >conditions. I could plug any numbers into any voting formula and you could >come up with a boundary condition that makes it look bad. That's true, but this particularly boundary condition sits right in the middle of the range of voting outcomes. >In practice, It's very unlikely (I can't think of a case) where >the voting is that low *and* that close. The vote for talk.politics.guns, posted last week, came in 122 to 20, as I recall; it would have failed your criteria, but passed if 8 more "No's" had been received. I'd rather opt for BC's that are truly at the boundaries. Fore example, how about "At least 100 more Yes than No votes, and Yes must comprise at least 67% of the vote." Then 122 to 20 passes; 99 to 0 fails; 200 to 50 passes, but 201 to 101 fails. Thus, for a "hot" group (one receiving lots of votes), it reduces to "67% of the vote must be "Yes"", while for a small group, it can never result in a situation where more No votes could have caused the vote to pass. For these numbers, the boundary lies at 200 Yes, 100 No; that vote meets both criteria. With fewer than 300 votes, all you have to worry about is 100 more Yes than No; with more, all you have to worry about is 67% Yes. Of course, the 100-vote margin and the 67% figure can be changed to taste; this will simply move the boundary. >>I've brought this up before, and gotten a rousing "yawn" in response. >>Why vote against a group ? I can only think of a few reasons: > >There is another reason, which you evidently haven't heard my lecture on >over the last eternity here on the net: > >5) You do not want an overly complex namespace. The reason you don't want an >overly complex namespace is because it breeds confusion in the naive user. >There are two problems with leaving moribund newsgroups around: >o A naive user will tend to post to that newsgroup assuming there's an > audience. Since the group is moribund, there is by definition a very small > audience, causing the user to get a non-existant or minimal response. You > end up with either an unhappy user or a posting that gets sent out multiple > times trying to find an audience, negatively impacting net volume. Hmm. I'll avoid reducing this argument to absurd levels (reduction ad absurdum, or whatever the rhetoricists call it), and simply ask this: what examples are there of newsgroups with such low volumes and readerships that this has occurred ? I know of none. Certainly, there are low-volume newsgroups; soc.misc at times runs for long periods with no postings; then occasionally flares up into a raging debate. I've never seen a posting stating "I posted this to soc.misc, but didn't get a response, so I'm trying here", but, of course, I don't read the whole net. There may be examples in the more obscure comp* groups, which I don't read. My hunch is that many people do what I do; subscribe to lots of low-volume groups, even if my interest is marginal. It's not much work to "n" the odd posting once a week, and occasionally it's something I want to respond to. If you have examples, I'll accept your reasoning, of course. >Think of a moribund newsgroup as USENET's appendix. I have no complaint about removing groups; I just have a higher tolerance than you for how "active" a group should be. ------------------------------ valuable coupon ------------------------------- Bill Thacker att!cbnews!wbt "C" combines the power of assembly language with the flexibility of assembly language. Disclaimer: Farg 'em if they can't take a joke ! ------------------------------- clip and save --------------------------------
ewiles@netxcom.UUCP (Edwin Wiles) (03/28/89)
In article <3007@epimass.EPI.COM> jbuck@epimass.EPI.COM (Joe Buck) writes: >However, we've never had a newsgroup deletion rule before, and I have >one major concern about Greg's suggestion. Since a majority of 100 >"yes" over "no" votes is required to KEEP a group, I'm concerned about >the possibility of "vote harrassment" -- essentially frivolous votes on >smaller groups where the people that benefit from that group have to keep >voting again and again. How's this? You require a 6 month period between removal calls. That way, they only have to support it twice a year. Kinda like PBS. Unfortunately, this *will* require a central authority to keep track of when the last call was! Enjoy! -- ...!hadron\ "Who?... Me?... WHAT opinions?!?" | Edwin Wiles ...!sundc\ Schedule: (n.) An ever changing | NetExpress Comm., Inc. ...!pyrdc\ nightmare. | 1953 Gallows Rd. Suite 300 ...!uunet!netxcom!ewiles | Vienna, VA 22180
david@ms.uky.edu (David Herron -- One of the vertebrae) (03/28/89)
Chuqui -- how 'bout:
6) You don't want an overly simple namespace. The reasons you don't
want an overly simple namespace are twofold:
a) The naive user with his/her myriad of questions keeps finding
one place to post questions to when the questions are really on
many topics. This raises volume in each group needlessly.
b) Splitting traffic flow over a smaller number of groups makes
it easier for individual postings to be lost in the shuffle.
Oh yeah, small number of groups doesn't necessarily mean a simple
namespace, and neither does a large number of groups mean a complex one.
But they tend to follow from each other.
It's a balancing act between keeping the namespace simple & keeping
it usable.
--
<- David Herron; an MMDF guy <david@ms.uky.edu>
<- ska: David le casse\*' {rutgers,uunet}!ukma!david, david@UKMA.BITNET
<-
<- The problem with mnemonics is they mean different things to different people.chuq@Apple.COM (Chuq Von Rospach) (03/28/89)
>Chuqui -- how 'bout: >6) You don't want an overly simple namespace. The reasons you don't >want an overly simple namespace are twofold: >a) The naive user with his/her myriad of questions keeps finding > one place to post questions to when the questions are really on > many topics. This raises volume in each group needlessly. >b) Splitting traffic flow over a smaller number of groups makes > it easier for individual postings to be lost in the shuffle. Nope. You don't want it overly simple. You want it efficient. I'm arguing for efficient, not simple. To explain the difference, look at theoretical ideals for the two: o simple: the ideal simple network structure is a single large group. That way you never have cross-posted or incorrectly posted messages. o efficient: an ideal efficient name space means that every posting has a single, optimal newsgroup for posting and that there are no newsgroups that have no traffic. Everything gets used, and everything that's needed is there. Obviously, because the net is a constantly mutating beast, the ideal will never be met. But my feeling is that we can do some judicious pruning to make the net more efficient and reduce name-space confusion without doing major surgery or causing problems. In fact, if you take my definition at face value, a pruning that causes problems reduces efficiency and is therefore wrong. If I were for simplicity for simplicity sake, I wouldn't have been the person who fostered both comp.sys.mac.programmer and comp.sys.mac.hypercard. That doesn't increase simplicity. I do feel, however, that the groups increase efficiency -- because, "just say 'n' fanatics" to the contrary, there is a limit to the amount of volume you can wedge in a group before the group starts to collapse in onto itself because of the noise. >Oh yeah, small number of groups doesn't necessarily mean a simple >namespace, and neither does a large number of groups mean a complex one. >But they tend to follow from each other. Nope. You're right again. You don't create groups for the sake of creating, you don't trim for the sake of trimming. (inspired readers will notice that I'm not giving recommendations of group names that I'd recommend for pruning under this proposal. This is not accidental). Chuq Von Rospach -*- Editor,OtherRealms -*- Member SFWA chuq@apple.com -*- CI$: 73317,635 -*- Delphi: CHUQ -*- Applelink: CHUQ [This is myself speaking. No company can control my thoughts.] USENET: N. A self-replicating phage engineered by the phone company to cause computers to spend large amounts of their owners budget on modem charges.
richard@gryphon.COM (Richard Sexton) (03/28/89)
In article <27870@apple.Apple.COM> chuq@Apple.COM (Chuq Von Rospach) writes: > >5) You do not want an overly complex namespace. The reason you don't want an >overly complex namespace is because it breeds confusion in the naive user. >There are two problems with leaving moribund newsgroups around: This is the goal of the net ? To not confuse the naive user ? >o A naive user will tend to post to that newsgroup assuming there's an > audience. Since the group is moribund, there is by definition a very small > audience, causing the user to get a non-existant or minimal response. You > end up with either an unhappy user or a posting that gets sent out multiple > times trying to find an audience, negatively impacting net volume. >o The larger the namespace, the harder it is to figure out what the > appropriate newsgroup to post is. In many cases, it's ambiguous what groups > a posting belongs. In an efficient namespace, there would be a single > appropriate newsgroup for every posting, and cross-posting would be > unneccessary. USENET is anything but efficient, and perfect efficiency is an > unattainable goal anyway, but we should tweak the namespace towards > efficiency where reasonable. This includes both creating newsgroups to > handle subjects that aren't covered *and* deleting newsgroups that no longer > have a purpose. The net has a way of educating users. -- Keep out of the reach of children richard@gryphon.COM decwrl!gryphon!richard gryphon!richard@elroy.jpl.NASA.GOV
chuq@Apple.COM (Chuq Von Rospach) (03/28/89)
>The vote for talk.politics.guns, posted last week, came in 122 to 20, as >I recall; it would have failed your criteria, but passed if 8 more "No's" >had been received. I'd rather opt for BC's that are truly at the boundaries. Trust the net to toss in a test case when I'm not looking... >Fore example, how about "At least 100 more Yes than No votes, and Yes must >comprise at least 67% of the vote." Then 122 to 20 passes; 99 to 0 fails; >200 to 50 passes, but 201 to 101 fails. Thus, for a "hot" group (one >receiving lots of votes), it reduces to "67% of the vote must be "Yes"", >while for a small group, it can never result in a situation where more >No votes could have caused the vote to pass. I could live with this, happily. It does what I want: show a general interest for the creation of the group and show that this interest isn't limited to a rather small but vocal special interest contigent. Hmm. I rather like this -- better than my proposal, in fact. >Of course, the 100-vote margin and the 67% figure can be changed to taste; >this will simply move the boundary. 100 votes is traditional on USENET as a minimal acceptance level. 67% is fairly standard in American Government. Seem like reasonable numbers to me. (50%+1 is simply too easy, especially with lots of voting going on. I prefer a bias towards the standard quo. Imagine how things would change in Congress if everything requires 67%?) >There may be >examples in the more obscure comp* groups, which I don't read. yeah. The one I can think of offhand is comp.edu. I've seen a few postings get reposted to wider audiences from there. I haven't tracked USENET that closely over the last year, so I don't have what I'd be willing to call suggestions to fit the rule. Maybe there are none currently -- I just feel that based on my time on USENET it's a good thing to do, when it's necessary. That doesn't imply it currrently is. Chuq Von Rospach -*- Editor,OtherRealms -*- Member SFWA chuq@apple.com -*- CI$: 73317,635 -*- Delphi: CHUQ -*- Applelink: CHUQ [This is myself speaking. No company can control my thoughts.] USENET: N. A self-replicating phage engineered by the phone company to cause computers to spend large amounts of their owners budget on modem charges.
msb@sq.com (Mark Brader) (03/28/89)
I comment on three aspects of the guideline proposals. First: > o A group creation vote succeeds if 67% of the vote is for creation. > o A group deletion vote succeeds if 67% of the vote is for deletion. > o Any vote with less than 150 total votes will be considered failed for > lack of interest. I would support the above proposal, for much the same reasons as those advanced previously. However, I would like it even better if it read as follows: o A group creation vote succeeds if 2/3 of the vote is for creation. o A group deletion vote succeeds if 2/3 of the vote is for deletion. o Any vote where neither side reaches 150 votes will be considered failed for lack of interest. The change from 67% to 2/3 is simply a preference for small numbers; 2/3 is a common requirement for "important votes" in various bodies, but I don't think I've ever heard of 67% being used. The change in the wording of the rule I list third is because in the first version a vote of 130-0 is a loss but 130-50 is a win, and it doesn't seem fair to me that voting against a proposal can help it win. If the first version was adopted, people who thought a proposal was silly might be afraid to vote against it for fear of triggering this effect. (No matter how silly a proposal there is always SOMEONE who will vote for it!) The exact number 100 or 150 is not important, but I can't see any other reasonable way of establishing a threshold. I propose that 150 be adopted for now, but subject to annual review. Indeed, the entire guidelines should be subject to annual review. (The original proposal also had a rule about spoiled ballots and auditing; I have no comment on this, either way, and excluded it from this article for brevity. I do agree with most of the proposals I've seen for ensuring that individual votes come from real individuals.) Second: Back when voting was first adopted, there was no rule that a discussion period had to precede the 30-day vote period. So discussion and voting used to be held simultaneously. I think this was the principal reason that the period is as *long* as 30 days. And it is because the period is so long that "repeated calls for votes" are necessary, or at least, that some people think they are necessary. But how do people vote in, say, directors' meetings? There is a discussion period, and then someone decides that enough time has passed, and then there is a call for a vote and the result is tallied at once. It may not be fair to do exactly this on the net, because of net-people taking vacations, but I think we should move more in that direction, and should also try to reduce the total time required. I propose that there should be three phases. o First, 2 weeks or more of discussion, by the end of which time the time the question(s) to be voted on should be precisely established. o Second, 2 weeks in which discussion may continue and votes will also be accepted. During this time the proposer would be allowed, perhaps after seeing the early voting, to withdraw the proposal and perhaps start over at the first step with a modified one. Minor revisions (e.g. name changes expected to be non-controversial) would also be allowed during this period. o Third, 1 week during which votes would be accepted and voters would be allowed to change their original votes now that the entire discussion period is over. People who take vacations of more than 3 weeks are out of luck, but then, who has time to catch up on *any* group after *any* vacation these days? And if the proposal is good it would probably pass even if a small number of people didn't get to vote for it. Third, note that in my second set of proposals above I said "question(s)". While the original proposed new guidelines demanded that a vote should be on yes-or-no questions only, I see no reason to demand this, and reasons not to demand it. The requirement should only be that the questions are well-defined and are not phrased on such a way as to "split the vote" on the main question which should always be "create/delete or not". For example, this should be okay: Choose one: [ ] Do Not Create [ ] Create If you chose Create, also choose a name: [ ] No opinion [ ] comp.text.sq [ ] comp.text.sq.sqtroff [ ] comp.text.sqtroff [ ] comp.text.sqps [ ] Other: _________________ "Australian Rules" or similar preference voting should also be okay, if the proposer is willing to tabulate it! This would be appropriate for those votes where there is controversy over naming or over how many groups to create. For example: Rank the following in preference order, 1 = most preferred: [ ] Do Not Create [ ] Create comp.text.sq Only [ ] Create comp.text.sqps Only [ ] Create comp.text.sqtroff Only [ ] Create comp.text.sq.sqtroff Only [ ] Create Both comp.text.sq and comp.text.sq.sqtroff [ ] Create one group, other name: _______________ Either of these multi-way vote types should be permitted as long as: o There is consensus during the discussion period that the questions are fairly expressed, with all well-supported or well-conceived choices included; and vote-splitting is not being attempted. o The primary question is always Do Something or Do Nothing; 2/3 of the vote (or whatever guideline is finally chosen) is required to Do Something. This is required for *each* newsgroup proposed; in the last example, 2/3 of the votes would have to be for Create Both for that to pass. (The "product placement" in the examples was intended in a spirit of fun.) Mark Brader, SoftQuad Inc., Toronto, utzoo!sq!msb, msb@sq.com We can design a system that's proof against accident and stupidity; but we CAN'T design one that's proof against deliberate malice. -- a spaceship designer in Arthur C. Clarke's "2001: A Space Odyssey"
wisner@shadooby.cc.umich.edu (Bill Wisner) (03/28/89)
>The net has a way of educating users.
Your way of educating users is exactly what we're trying to change, Richard.pete@octopus.UUCP (Pete Holzmann) (03/28/89)
In article <27934@apple.Apple.COM> chuq@Apple.COM (Chuq Von Rospach) writes: >o efficient: an ideal efficient name space means that every posting has a > single, optimal newsgroup for posting and that there are no newsgroups > that have no traffic. Everything gets used, and everything that's needed > is there. You forgot one more attribute of the ideal efficient name space: the number of postings appearing in the newsgroup per <standard reading interval; maybe a day> is reasonable. There is a point where newsreading software just can't handle posting volume efficiently. When I come in on a monday and find 200 new articles in c.s.i.p, I am not happy. The postings may all "belong" there, but I wish there were more than one group covering the general topic of PC's. It would be more efficient. Pete -- Peter Holzmann, Octopus Enterprises |(if you're a techie Christian & are 19611 La Mar Ct., Cupertino, CA 95014 |interested in helping w/ the Great UUCP: {hpda,pyramid}!octopus!pete |Commission, email dsa-contact@octopus) DSA office ans mach=408/996-7746;Work (SLP) voice=408/985-7400,FAX=408/985-0859
dan@ccnysci.UUCP (Dan Schlitt) (03/28/89)
In article <37740@bbn.COM> fwebb@BBN.COM (Fred Webb) writes: > >Greg, if you're willing to take on this thankless job, more power to you! > >In general, I think your proposed guidelines are fine, and are pretty close >to what's been happening. I do have a couple of minor comments. I agree with both these quoted statements. You're a brave person, Greg. One thing that needs to be added for clarity is the parts of the news tree that the rules apply to. Clearly they don't apply to alt. That was intended to be chaos. If the alt people want to change things they are the ones to do it. There was a comment about bionet in one article. Again, it would be my assumption that these rules don't apply there. That is Elliot Lear's domain. But the rules need to be clear on this. On the subject of group creation, the "rule of 100" seems to work reasonably well. I don't see any real reason to fool with it in the way that some folk are suggesting. As I understand it, the number was derived as a cost crossover between mailinglist and newsgroup. It seems to me that this is still the relevant consideration. I'm not sure how this crossover scales with the size of the net. My first inclination is that it is not very sensitive. Group removal is something new. The comment that it should be somewhat harder to take people's newsgroups away than it was to create it in the first place has my sympathy. The newsgroup mailinglist crossover would again be the relevant measure. Lack of traffic should be the trigger and not the noise created by some group of netters. -- Dan Schlitt Manager, Science Division Computer Facility dan@ccnysci City College of New York dan@ccnysci.bitnet New York, NY 10031 (212)690-6868
davidbe@sco.COM (The Cat in the Hat) (03/29/89)
compton@silver.UUCP (Dave Compton) said:
-
- I have another suggestion. Either a vote > 100 vote margin,
-or 250 or more yes votes. Say there is a call for votes for a
-newsgroup with a controversial subject. And for the sake of argument,
-the vote come out to.
-
-Yes 300
-No 299
-
- Now, In my humble opinion, this newsgroup has as much right to
-exist as a speciality group that passes with the vote of:
-
-Yes 105
-No 4
I disagree. The fact that almost 300 people think that a newsgroup is a
bad idea should be a clue that, perhaps, something is wrong. At any
rate, the vote should be re-discussed, and then voted on at a later (quite
a bit later) time.
--
David Bedno (The Cat in the Hat), Reachable at: davidbe@sco.COM -OR-
.!{uunet,sun,ucbvax!ucscc}!sco!davidbe -OR- At home: 408-425-5266
At work: 408-425-7222 x5123 (I'm probably here...)
Disclaimer: Speaking from SCO but not for SCO. Not by a long shot.
Performance is life. Entertainment is death.
- World Entertainment Wardavidbe@sco.COM (The Cat in the Hat) (03/29/89)
woods@ncar.UCAR.EDU (Greg Woods) said:
-
-
- RULES FOR USENET GROUP CREATION AND DELETION (draft)
[ etc, etc...Greg just formalizes the current situation. ]
[ I don't think anyone has any real objections with this.]
-
- RULES FOR DELETING A USENET GROUP
-
-[Since there has obviously been no consensus ever reached on this, this
- statement cannot be construed as anything other than my personal opinion]
-
-I propose that the procedure be the same as for group creation, with
-100 more keep than delete votes required to save the group, with the exception
-that all calls for votes and discussions MUST be cross-posted to both
-news.groups and the group in question. This will obviously only work for
-unmoderated groups; I am open to discussion on procedures for deleting
-moderated groups, but I would prefer to wait for a final resolution on
-the rec.humor.funny/Brad Templeton controversy before trying to formalize
-them, because it will be IMPOSSIBLE to keep the flames down until then.
I've got a couple of suggestions/clarifications here.
1) Same procedure as for group creation includes a discussion period before
a vote.
2) A vote should always require a margin [however much is decided on] to
CHANGE the status quo. Thus, there should be 100 more yes than no votes
to create a newsgroup, and 100 more delete votes than keep votes to remove
a group.
3) This rule should, of course, not apply to newsgroups with significant
amounts of volume or readership. Hell, forget significant; substitute
negligible. Removing a newsgroup should only be an extreme thing.
4) There is no number 4. Ponder the above 3 and see if it's so bad for you
to deal with.
--
David Bedno (The Cat in the Hat), Reachable at: davidbe@sco.COM -OR-
.!{uunet,sun,ucbvax!ucscc}!sco!davidbe -OR- At home: 408-425-5266
At work: 408-425-7222 x5123 (I'm probably here...)
Disclaimer: Speaking from SCO but not for SCO. Not by a long shot.
Performance is life. Entertainment is death.
- World Entertainment Wardavidbe@sco.COM (The Cat in the Hat) (03/29/89)
roy@phri.UUCP (Roy Smith) said:
-
- I pretty much agree with Greg Woods's proposed rules. I only have
-one comment; I think the "100 more Yes than No" rule is not strong enough.
-I don't remember exactly when the 100 threshold was thought of, but I'm sure
-the net was 1/Nth the size it is today, where N is in the interval (2,5).
Actually, it was only a little over a year ago that the 100 threshold was
thought of.
It actually makes sense when you look at how many of the votes have gone.
The number of newsreaders and sites may be constantly increasing, but the
number of people who actually seem to be participating in the actual
administration of the net (by discussing, voting, yelling about new groups)
is fairly constant. Most news readers are sheep (lurkers) and don't care
about what new groups get created. Or if they do, they don't say anything.
100 is still a good number.
Maybe 150 would be better for now, since it has more room for future
expansion, but there's nothing wrong with the way the rules are now. At
least, not anything more wrong than any of the other suggestions that have
been posted.
--
David Bedno (The Cat in the Hat), Reachable at: davidbe@sco.COM -OR-
.!{uunet,sun,ucbvax!ucscc}!sco!davidbe -OR- At home: 408-425-5266
At work: 408-425-7222 x5123 (I'm probably here...)
Disclaimer: Speaking from SCO but not for SCO. Not by a long shot.
Performance is life. Entertainment is death.
- World Entertainment Warrusty@cadnetix.COM (Rusty) (03/29/89)
In article <8281@csli.STANFORD.EDU> cphoenix@csli.stanford.edu (Chris Phoenix) writes: <<Sorry, I lost the attribution of the next quote. My apologies>> >>o In both cases, if the number of spoiled or proven missing ballots exceeds >> 2% of the total vote, or if enough spoiled/missing votes are found that >> would have changed the result of the vote, the vote must be re-taken >> by an agreed-upon neutral third party. > >Given the vagaries of e-mail, this seems a bit harsh. Far more than 2% >of my off-system letters get bounced. I'd be very surprised if many >votes would pass this test. > >... >Chris Phoenix >cphoenix@csli.stanford.edu An example. the vote for rec.ham-radio.rules failed by 5 votes. The vote was about 126 to 31 (or some such numbers, I'm going from memory). If 10% of the votes did not make it, and if the votes that did not make it were of the same ratio as the actual vote, then rec.ham-radio.rules did not get created *because* email failed to get the vote through. (And since the connectivity of the vote-taker was EXTREMELY poor, I can believe that the number of failed votes was much worse than 10%. I even seem to remember seeing at least 5 postings of votes (invalid, of course) FOR creation of the group... posted because they could not get through to any of the 3 paths suggested for voting.) Now I agree that email tends to be less than perfect. However, just because something normally works a certain way, does that mean that we should ignore its effect? Just because mail can be highly unreliable at times, does that mean that we penalize the proposed group just because the proposer happened to be poorly-connected? Had I realized how poorly connected the r.h-r.rules vote-taker was, *I* would have volunteered to be the vote-taker, as I seem to be pretty well-connected. (To all but gary delong, the r.h-r.rules vote-taker!) However, I did not know that there was going to be such a problem. And now, r.h-r.rules is not going to be created, and the volume in r.h-r will continue. (Sure, it has slowed down a bit at the moment, but just wait 'till the proposal which will get sent to the FCC gets announced!). Since the vote failed, I will NOT call for another vote (even though I bet it would pass!). However, it seems that, in the future, the vote guidlines should take into account the fact that mail is not nearly as reliable as we would like. I would suggest that the guideline go like this: If enough spoiled/missing votes are found that would have changed the result of the vote, then those votes should be counted and the result changed. (Uh-oh, here come the flames!) ---------- #include "quote.cute" Rusty Carruth UUCP:{uunet,boulder}!cadnetix!rusty DOMAIN: rusty@cadnetix.com Daisy/Cadnetix Corp. (303) 444-8075\ 5775 Flatiron Pkwy. \ Boulder, Co 80301 Radio: N7IKQ 'home': P.O.B. 461 \ Lafayette, CO 80026
dlm@cuuxb.ATT.COM (Dennis L. Mumaugh) (03/29/89)
In article <8280@csli.STANFORD.EDU> cphoenix@csli.stanford.edu (Chris Phoenix) writes: ... and I don't care about most of the stuff posted, so why should the size of the net affect whether or not a group gets created? A group now is the same as a group several years ago, no? If a group with 100 members was worthwhile back then, why should it change now? If you read the arbitron reports and the estimate of the cost of posting an article, you will see that there needs to be a line drawn. Each article takes a small amount of money for each site. Cost of phone calls, ammortized cost of equipment, etc. Also each article posted means the use of some resource -- a disk block, an inode, some cpu cyles, some modem bauds -- that could have been used for a better purpose. As more people start reading the net, the percentage of readers required to form a group will get smaller, and so there will be more groups. On the other hand, this hasn't killed us yet. If groups are required to have more readers, then they will have a higher volume, and some topics will be squeezed together. From what I've seen, high volume is much more of a problem for users than fragmented topics. (I've been forced to drop several groups already.) Information overload is the problem. We have no such problem in a sense as the amount of disk storage in the universe has a limit. In a practical sense, I have 128,000 blocks for netnews. But in reality if the volume goes up the residence time of an article goes down. Too short a residence time results in news discussions becoming volatile as people fail to realize that the exact same point was raised only ### days earlier. So the choice is between letting the net have more groups (annoyance for the administrators) and forcing the groups to gain more readers and posters (major problems for the users). IMHO, the users are more important. 1/2 :-) I made a coment earlier (six months ago) without any reaction on the net: USENET is NOT a substitute for the Dow Jones Ticker, the AP news wire and an up load of the typesetter tapes of all textbooks and magazines, etc. It cannot duiplicate the library. As such one should not have the completeness obsession: the namespace expands to include all possible subjects. The original plans for creation of a news group was that one was created ONLY after traffic proved it was viable. Newgroups served to manage existing traffic not generate new traffic. I still think we need to operate that way, otherwise everntually we will have a news group for every subject in the library and change from names to a Dewey Decimal system. -- =Dennis L. Mumaugh Lisle, IL ...!{att,lll-crg,attunix}!cuuxb!dlm OR dlm@cuuxb.att.com
sommar@enea.se (Erland Sommarskog) (03/29/89)
Chuq Von Rospach (chuq@Apple.COM) writes: >100 votes is traditional on USENET as a minimal acceptance level. 67% is >fairly standard in American Government. Seem like reasonable numbers to me. >(50%+1 is simply too easy, especially with lots of voting going on. I prefer >a bias towards the standard quo. Imagine how things would change in Congress >if everything requires 67%?) So what has the American Government to do with Usenet? OK, 2/3 majority is fairly common elsewhere to, when a qualified majority is required. As for the issue at hand, I see no reason to change the existing rules. -- Erland Sommarskog - ENEA Data, Stockholm - sommar@enea.se I used to say "It could have been worse, it could have been Pepsi", then I drank a Diet Coke...
dtynan@altos86.UUCP (Dermot Tynan) (03/29/89)
In article <1634@ncar.ucar.edu> woods@ncar.UCAR.EDU (Greg Woods) writes: > > RULES FOR DELETING A USENET GROUP > >I propose that the procedure be the same as for group creation, with >100 more keep than delete votes required to save the group, with the exception >that all calls for votes and discussions MUST be cross-posted to both >news.groups and the group in question. This will obviously only work for This is incorrect. The 'burden of proof' so to speak, is on the people who want the group *removed*. In reality, the only valid groups to come up for 'rmgroup'ing, will be those which are not special-purpose (such as n.a.important) or those which have very low activity (in which case, it is easy to obtain a 100-vote delete margin). If it's possible to 'rmgroup' without the 100-vote margin, then people will rely on apathy to rmgroup certain groups which are indeed active and legit. On the other hand, a margin in favor of deleting shows strong interest in seeing the group go, and the disk-space (?) freed. As for moderated groups, the same rules should apply. If the DECWRL neilsons show that the group is being read, and the moderator has no problem getting contributions, then why remove it? In fact, can we now stop the chatter about r.h.f, and assume that those who feel betrayed by the 'commercialism' will refrain from posting to the group, and those that do contribute will realise their material (which is rarely original to begin with!) may pass on to other sources, perhaps for monetary gain. Either that, or we create a news.gibberish! - Der -- Return Path: dtynan@altos86.UUCP {..!pyramid!altos86!dtynan} Or use: dtynan@zorba.Tynan.COM -+-+-+- May the blessings of Jeyes Fluid fall upon you -+-+-+-
cooke@gumby.cs.wisc.edu (David Cooke=) (04/03/89)
In article <11326@s.ms.uky.edu> david@ms.uky.edu (David Herron -- One of the vertebrae) writes: > >And in case anybody thinks that such high numbers of votes aren't >possible, all you need is one volatile subject (like comp.women) >to set it off and you'll get lots of votes. > >My feeling is that in such a case, even though a lot of people have >indicated favourably, that a *LOT* of people have *ALSO* indicated >unfavorably. And that group of people is large enough that they >should be given attention. That there must be something fundamentally >wrong with the proposal if it's generating that much negative feelings. >*REGARDLESS* of the amount of positive feelings. Why is it relevant that a lot of people have voted no? No one is *forced* to read a group. What gives a very vocal group the right to suppress the interests of others on the net? If there is sufficient interest in forming a new group, one that will be highly active and reduce the noise on other groups, what's the problem? There are a lot of groups on the net that I will never be interested in reading, but I don't dispute the rights of those who wish to to read them regularly. Dave Disclaimer: I don't do disclaimers.