🚀 go-pugleaf

In article <8324@lib.tmc.edu>, jmaynard@oac.hsc.uth.tmc.edu (Jay Maynard) writes:
|> In article <1993Jan04.213001.63448@watson.ibm.com> strom@Watson.Ibm.Com (Rob Strom) writes:

|> >While occasionally it could work the other way, I think
|> >the benefits of (1) guaranteeing subject lines which are
|> >both clear and short and (2) avoiding situations like
|> >the rms vote outweigh any deficiencies. 
|> 
|> If you do actually achieve 1, then it's a Good Thing. Avoiding 2 is nearly
|> impossible, though; there will always be a sore loser, and even occasionally a
|> sore winner (Hi, Joel).
|> 

I can't avoid *all* situations like the rms vote, but
even avoiding a few is worth it.


|> >Do we have logs of past votes?  Can we determine what
|> >fraction of the last, say, 100 renaming ballots would
|> >have had to be split under my proposed rules?
|> 
|> The data is available from ftp.uu.net, I think, in something like
|> usenet/news.announce.newgroups.

I followed your advice.  Site ftp.uu.net indeed has about a year's
backlog of votes/proposals.  They're not categorized as multiple or single,
but using the heuristic that multiples tend to be tagged either -reorg or -newgr,
I brought up and scanned all the logged groups having these names.

Only 22 multiples having the tag -reorg or -newgr were found,

Of these 22, only two (!) would have had to be split according
to my proposed criteria.

These were:
  (1) news.admin-reorg, which simultaneously renamed
  news.admin and news.sysadmin

and, of course

  (2) rec.music.makers-newgr, which as everyone knows by now,
  simultaneously renamed alt.guitar* and rec.music.synth.

That's one in eleven.

Of the two affected reorgs, at least one (the rmm*) was
sufficiently problematical that splitting the ballot
would have avoided problems.

The vast majority of reorgs just took a single existing group
of the form a.b, renamed it a.b.misc, created a.b.<various others>,
and narrowed the charter of a.b.misc to exclude discussions
of topics covered by the various others created.
These simple reorgs would be unaffected by my proposed changes.

Oh by the way, it is easy to find the smoking pistols
John requested:  that is, CFVs which explicitly redefine
the charter.  (Although most groups which created a misc
clearly implied that topics covered by other groups would
be excluded from misc, not all made it totally explicit.)
Here's an example from the rec.bicycles* reorg:

>It is further proposed that the existing rec.bicycles be renamed and
>redefined as follows.
 ^^^^^^^^^
>rec.bicycle.misc:  General riding techniques, rider physiology,
>injuries and treatment, diets, and other cycling topics.

My sampling may be skewed since there may be multiple rename
votes which were archived under names which didn't make that clear.

But I think the above is indicative of the minimal negative
effect splitting the balloting would have had.

For the cases where a vote is split, I think we can make
the vote-taker's job easy enough by having the votes mailed
to a common address and using a form which would facilitate
automatically merging the multiple votes, if needed.

I still need to work on refining the proposal to clarify
exactly which new group creations ought to be on a common
ballot with renames.  


-- 
Rob Strom, strom@watson.ibm.com, (914) 784-7641
IBM Research, 30 Saw Mill River Road, P.O. Box 704, Yorktown Heights, NY  10598