bernstei@hpuxa.ircc.ohio-state.edu (Dan Bernstein) (10/22/89)
The time has come to pretend that this is a real issue, to be discussed
and voted upon like everything else in news.groups. (Should readers of
news.announce.newgroups hear about this? Greg?)
How about the following changes:
1. Near the end of the discussion period, before posting a call for
votes, whoever is planning to tally votes should post a NAME CHECK
to news.announce.newgroups. The NAME CHECK should list all names
proposed for the new group during the discussion period, in
alphabetical order. This means that the name need not be settled
during the discussion period, though settling it would be nice.
2. The vote taker should accept mailed requests to add a name to the
name list. At least three days after the initial NAME CHECK is
posted, the vote taker should post his CALL FOR VOTES in
news.announce.newgroups; the CALL FOR VOTES should list all
proposed names for the new group, including those submitted by
mail, in alphabetical order.
3. Voters should list some or all of the proposed names, each
accompanied by a YES, blank, or NO. YES and NO may have a
number attached, indicating a ``weight''; the weight is 1 by
default. A weight higher than 1 doesn't make that voter more
important; it deemphasizes the names with smaller weights.
4. To tally a single vote, the vote taker should divide the weight
of the YESs by the maximum YES weight, and similarly for the NOs.
(In other words, each vote can contribute any fraction between
1NO and 1YES to each name.) Then he should add up all the YESs
for each group, and add up all the NOs for each proposed name,
taking fractions into account. The vote taker can ignore write-in
names---that was the purpose of the NAME CHECK.
5. Finally, the vote taker should subtract the total NOs from the
total YESs, separately for each name; if the name with the highest
margin has that margin 100 or greater, the group passes with that
name. Otherwise, the group fails.
6. The CALL FOR VOTES should announce the above rules.
7. (This is a big jump:) If the question of moderation versus
nonmoderation is not settled during the discussion period, it
should be handled just like the question of names.
Thoughts:
A. This is a general form of approval voting (that's the name used by
the Mathematical Association of America for their successful version),
STV, the Australian system, the vote-often system, Dan's system, or
whatever you want to call it. The weights eliminate the failures of
the other variants that have been discussed. For example, if you want to
indicate that you approve of rec.aquaria, rec.fish, and rec.aquarium,
are neutral towards comp.society.aquaria, and dislike sci.aquaria:
rec.aquaria YES
sci.aquaria NO
comp.society.aquaria (this line could be left out)
rec.fish YES
rec.aquarium YES
If you really prefer rec.aquaria and only feel slightly inclined
towards rec.fish, and consider comp.society.aquaria a lesser evil
than sci.aquaria:
rec.aquarium YES3 (YES^3, 3 YES, YES YES YES, or whatever)
rec.aquaria YES4
rec.fish YES
comp.society.aquaria NO
sci.aquaria NO 5
And so on. The idea is that the voters can precisely express how
happy they'd be with each name.
B. What do people think of the moderation idea? For example:
rec.humor.funny (mod Brad) YES
rec.humor.funny (unmod) NO
rec.humor.stupid (mod Shmoe) NO
rec.humor.stupid (unmod) NO
C. There should be some way to deal with idiots who submit a hundred
different names for a group, and shout ``illegal vote!'' if the
vote taker ignores them. Will this be a problem? Also, is three days
long enough to make sure that write-in names get in?
D. I don't think wildcards are worth the fuss: it doesn't take that long
to write ten YESs or NOs, and the main argument for wildcards---to
support write-ins that you haven't heard of yet---is moot. Do people
agree that we don't need wildcard votes?
E. In case the rules change from, e.g., margin to percentage, this
system stays intact.
F. Should there be a way to vote, e.g., with all the YESs having weights
under 1? How often are people only slightly satisfied with any of
the names---but satisfied enough to say YES? Does that even make
sense?
F. Is this simple enough for the typical novice voter to understand?
In a few weeks, if all goes well, somebody should post a call for votes
on this issue, with the explicit reminder that the votes are for
information only. Of course, this isn't really a new group issue, and
the guidelines don't really apply, so after we get results we'll have to
sit back and think about what they mean. (What a concept.)
By the way, let's keep this discussion in news.groups, not news.admin.
---Dan Bernstein, brnstnd@acf10.nyu.edu, bernstei@hpuxa.ircc.ohio-state.edujgreely@oz.cis.ohio-state.edu (J Greely) (10/22/89)
Static host tables, flat user lists, and now Robert's Rules of net.order. By god, the man *is* a bureaucrat. Nice touch making a usa-only call for discussion on a netwide issue, too. Surely the folks across the pond couldn't possibly be affected by your proposed changes. "I don't like his filed teeth and his watery, yellow eyeballs!" "You can't even see his face from here!" "I don't like his belligerent elbows, his threatening ear-backs!" -=- J Greely (jgreely@cis.ohio-state.edu; osu-cis!jgreely)
peter@ficc.uu.net (Peter da Silva) (10/23/89)
First of all, this proposal is *not* a variant of the STV system. This is a weighting system, whereas STV is a priority system. If you don't under- stand the difference between weighting and priority, don't worry... you're in good company: the people who designed the UNIX scheduler didn't either. Trust me... there's a difference. Secondly, this is the most complex proposal I've seen so far, with no clear advantage over STV or even Alien Wells' scheme. There's a problem that needs to be fixed, but anything this complex is just not going to work. -- Peter da Silva, *NIX support guy @ Ferranti International Controls Corporation. Biz: peter@ficc.uu.net, +1 713 274 5180. Fun: peter@sugar.hackercorp.com. `-_-' "I feared that the committee would decide to go with their previous 'U` decision unless I credibly pulled a full tantrum." -- dmr@alice.UUCP
bernstei@hpuxa.ircc.ohio-state.edu (Dan Bernstein) (10/23/89)
In article <6629@ficc.uu.net> peter@ficc.uu.net (Peter da Silva) writes: > Secondly, this is the most complex proposal I've seen so far, with no clear > advantage over STV or even Alien Wells' scheme. There's a problem that needs > to be fixed, but anything this complex is just not going to work. For practical purposes, this is the same as what Wells proposed---with the advantage that *if* a voter wants to make fine distinctions between his YESs and NOs, he can. Most votes will look the same as votes under Wells' scheme. There's going to be some complexity in the new voting rules, no matter what. ---Dan Bernstein, bernstei@hpuxa.ircc.ohio-state.edu
peter@ficc.uu.net (Peter da Silva) (10/23/89)
In article <362@nisca.ircc.ohio-state.edu> bernstei@hpuxa.ircc.ohio-state.edu (Dan Bernstein) writes: > There's going to be some complexity in the new voting rules, no matter what. There doesn't need to be quite this much. And adding fractional votes significantly complicates the counting procedure. Really, the STV system is simpler than the one you proposed. -- Peter da Silva, *NIX support guy @ Ferranti International Controls Corporation. Biz: peter@ficc.uu.net, +1 713 274 5180. Fun: peter@sugar.hackercorp.com. `-_-' "I feared that the committee would decide to go with their previous 'U` decision unless I credibly pulled a full tantrum." -- dmr@alice.UUCP
allbery@NCoast.ORG (Brandon S. Allbery) (10/24/89)
As quoted from <361@nisca.ircc.ohio-state.edu> by bernstei@hpuxa.ircc.ohio-state.edu (Dan Bernstein): +--------------- | How about the following changes: +--------------- The system is considerably more complex than the current one, and people seem to have enough trouble following *those* rules at times. +--------------- | name list. At least three days after the initial NAME CHECK is +--------------- Not everyone is capable of, or runs, NNTP. At absolute MINIMUM it should be a week; two weeks is better. +--------------- | 5. Finally, the vote taker should subtract the total NOs from the | total YESs, separately for each name; if the name with the highest | margin has that margin 100 or greater, the group passes with that | name. Otherwise, the group fails. +--------------- This is susceptible to divide-and-conquer, whereby enough yes votes exist to create the group but they are divided over a number of names which individually do *not* have enough votes to succeed. This means that a group for a given topic might fail despite a clear majority for the group. +--------------- | 7. (This is a big jump:) If the question of moderation versus | nonmoderation is not settled during the discussion period, it | should be handled just like the question of names. +--------------- Were there a *simple* way to handle it, quite possibly. As is, no. Why not just hold a pre-vote on name and/or moderation status? +--------------- | rec.aquarium YES3 (YES^3, 3 YES, YES YES YES, or whatever) +--------------- AARGH!!! Imagine counting 1000 votes *by hand* because each voter used her own variant of this. +--------------- | And so on. The idea is that the voters can precisely express how | happy they'd be with each name. +--------------- It doesn't really cover the case where I vote against a name but don't want to vote for any specific group name too well. In particular, a voting system allowing any of {1yes/1no, 0yes/1no, 1yes/0no} is less than fair. How about a 50/50 system? (and the complexity increases...): 0.5yes/0.5no instead of 1yes/1no in the above summary. Still, correcting this glitch adds even *more* complexity to a system which is too complex already. +--------------- | C. There should be some way to deal with idiots who submit a hundred | different names for a group, and shout ``illegal vote!'' if the | vote taker ignores them. Will this be a problem? Also, is three days | long enough to make sure that write-in names get in? +--------------- (1) 1/100th of a vote isn't likely to count for much either way. (2) If she submits 100 votes, it's a pretty good chance that 90 or more are write-ins. You *did* say that write-ins could be ignored.... As stated above, three days is NOT long enough. You've just disenfranchised the UUCP contingent. +--------------- | D. I don't think wildcards are worth the fuss: it doesn't take that long +--------------- Wildcards are yet more complexity. Forget 'em. +--------------- | F. Should there be a way to vote, e.g., with all the YESs having weights | under 1? How often are people only slightly satisfied with any of | the names---but satisfied enough to say YES? Does that even make | sense? +--------------- Hoooooooo boy. Shall we release a 1000-page explanation of the Usenet Voting Guidelines after implementing all of this? +--------------- | F. Is this simple enough for the typical novice voter to understand? +--------------- H*ll, you've got *my* head spinning. Too complex and you disenfranchise the novices as well. (Although I daresay it might be argued by some that they ought to be disenfranchised until they learn the ropes.) Capsule summary: too complex. ++Brandon -- Brandon S. Allbery, moderator of comp.sources.misc allbery@NCoast.ORG uunet!hal.cwru.edu!ncoast!allbery ncoast!allbery@hal.cwru.edu bsa@telotech.uucp 161-7070 (MCI), ALLBERY (Delphi), B.ALLBERY (GEnie), comp-sources-misc@backbone [comp.sources.misc-related mail should go ONLY to comp-sources-misc@<backbone>] *Third party vote-collection service: send mail to allbery@uunet.uu.net (ONLY)*
woods@ncar.ucar.edu (Greg Woods) (10/25/89)
If it takes a 79-line article to describe the proposed rules, they are too complicated. It won't work in practice no matter how nice it sounds. --Greg
karen@everexn.uucp (Karen Valentino) (10/26/89)
bernstei@hpuxa.ircc.ohio-state.edu (Dan Bernstein) writes: >How about the following changes: > 2. The vote taker should accept mailed requests to add a name to the > name list. At least three days after the initial NAME CHECK is > posted, the vote taker should post his CALL FOR VOTES in > news.announce.newgroups; the CALL FOR VOTES should list all > proposed names for the new group, including those submitted by > mail, in alphabetical order. Let's make this closer to a week. If a group is worth having, it's worth waiting for.....Not all of us read news *every* single day (sacrilege, I realize, but it had to be said). I'd hate to miss out on my opportunity to contribute the name that says it *all*, just because there was other stuff I simply had to get done. [Lots of sensible but very intricate voting procedure stuff deleted.] >C. There should be some way to deal with idiots who submit a hundred > different names for a group, and shout ``illegal vote!'' if the > vote taker ignores them. Will this be a problem? Also, is three days > long enough to make sure that write-in names get in? I'm no expert, but I've seen enough merry prankstering on the net that I find it hard to believe that this wouldn't be a *real* problem, and a real headache for the poor vote taker. (On the other hand, if you want to keep the number of groups to a minimum, this could be part of a very strenuous "running of the gauntlet" that a group champion would have to endure in order to get his beloved group up and running. :-) ) Think about it: If you were championing a group about, say, aquaria life as we know it, would you want to deal with (and inflict on everyone else by insisting that they vote on each name separately) 50 names such as rec.goldfish.arent.they.cute because more than one person out there might want to mess things up for you? It would make sense that only names that were contributed with honest sincerity, with no malice aforethought, would be considered "real" contributions--but who's going to decide that? The person who is initiating everything has his/her own agenda (and why not?). I have no answers to this issue; questions and skepticism aplenty. >F. Is this simple enough for the typical novice voter to understand? I can speak to this one, because I'm a typical novice voter. I had to go over this article several times, carefully, to feel fairly sure that I understand the ins and outs of this system. Simple, it ain't. I'm wondering if the question stated here shouldn't be, "Is this simple enough for the typical novice voter to *use*?" Asking the voter to be able to figure things out enough to vote is a lot different than asking him/her to understand the voting system. As it is, you will have to provide an article that is available to the novice voter, day or night, that lays out the complexities stated above, if you want him/her to understand the voting system. In addition, I'd suggest that there be another article which would be the "HowTo" article, that would instruct the voter how to vote, in as linear a fashion as possible ("First, you do this. Next, do this. Next,...). That way, your typical novice voter who really couldn't care less how the voting system works could just get the damn vote in and get back to his/her newsreading. Karen -- Karen Valentino <> Everex North (Everex Systems) <> Sebastopol, CA ..pacbell!mslbrb!everexn!karen "Buy land. They've stopped making it." -- Mark Twain (thanks to jeffd for the loan of this quote)