Subject:
|
Re: CFV Clarifications and REVISIONS
|
Newsgroups:
|
lugnet.admin.general
|
Date:
|
Wed, 18 Aug 1999 20:43:00 GMT
|
Viewed:
|
205 times
|
| |
|
|
In lugnet.admin.general, "Adam Howard" <abhoward10@hotmail.com> writes:
> > I don't think that follows. It certainly wasn't specified in the CFV
> > that the votes would be interpreted in a cascading way. Note that (a)
> > says "All lugnet.* groups," it doesn't say, "As many lugnet.* groups
> > as possible," and it doesn't say that a vote for (a) implies a vote for
> > (b) as well.
>
> No it doesn't state that in the CFV, but it is clearly stated (IMO).
If everyone think that a, b, and c are mutually exclusive choices, yes, then
it's clearly stated -- except it's not clear whether a would cascade into b;
it doesn't specify that. It doesn't specify whether voting for a also
implies voting for b, or whether voting for a is an all-or-nothing vote.
> Perhaps it's just my understanding that lugnet.* means Every single group on
> LUGNET including all the lugnet.cad.* groups.
Certainly it does -- I don't there's any arguing that point.
> I was under the impression
> that everyone voting would understand that .* means everything under that
> heading.
What do you mean by "everything under"? All of the sub-groups expanded, or
all of the vote choiced under? (Or both?) It's ambiguously worded because
it leaves that part out.
> Therefore: If you vote YES to a and b and list something in c and vote
> YES to that too, and you vote YES to 2, then you have cast 1 single
> vote.
What do you mean by one single vote? One multi-part vote? Or one single
vote as opposed to multiple votes (where multiple votes would count more
than a single vote, which hopefully isn't the case here).
> That one single vote goes towards all lugnet.* groups (including cad
> and any other groups listed by other voters in c).
>
> Lets say 100 people vote (and all voted YES to 2)
>
> a = 50 vote YES
> b = (50) + 20 = 70 vote YES 2/3-majority b passes
> c = (50) + 30 = 80 vote YES 2/3-majority whichever individual c('s)
> pass(es)
It doesn't state in the CFV that a would cascade into b and/or c. If you
want to interpret it that way, then you should make sure that all the voters
understand this interpretation, so they can re-cast their votes if
necessary.
Were I to have voted, I probably would have cast my vote thusly:
a - YES
b - NO
c - NO
But under your interpretation, it sounds like I would have wanted to have
cast it as either
a - YES
b - (blank)
c - (blank)
or
a - YES
b - YES
c - YES
It's not clear to me if and how a, b, and c interrelate.
> b and c don't have a superficial majority becaus 50 people did vote for the
> groups listed in b and c when they voted YES for a. 70 people could have
> voted for a and then a would have passed too.
Well, certainly, if a passes, then b and c are moot. If a doesn't pass,
then there's an interpretation problem because of the ambiguity -- because
it doesn't say in the CFV anything about mutual exclusion or cascading.
> > What do you do about someone who expected they were casting a vote in
> > favor of all or nothing? (That is, only a complete ban and not a partial
> > ban.)
>
> If they vote YES to lugnet.*=ALL (a) and 2 YES, and it turns out that there
> is only a 2/3-majority for either b or c, a partial ban, then what you are
> saying is they would want to change their vote to NO for 1 and 2? Why on
> earth would they want to do that? That would probably cause the partial ban
> to fail because it would remove their vote from all the b and c categories.
I'm not saying anyone should necessarily change their vote. I'm saying that
the CFV is ambiguous as to how a, b, and c interrelate. It leaves a lot up
to the voter to either guess or assume or ask, and that's not a sound
method. It shouldn't leave anything at all up to guesswork or assumptions.
It's also not fully clear how 1 and 2 interrelate. When tallying the votes
for 2, does the vote tallier plan to apply the results of 1(a,b,c) only to
YES-votes in 2 and to ignore the results of 1(a,b,c) for NO-votes in 2?
(That's what I would guess, but it's not stated.)
As currently worded, the only way any kind of ban can occur with this CFV is
if:
i) the total number of people who voted YES for 2 and who also voted YES
for 1a equals or exceeds 2/3 of the total number of people voting,
or
ii) the total number of people who voted YES for 2 and who also voted YES
for 1b equals or exceeds 2/3 of the total number of people voting,
or
iii) the total number of people who voted YES for 2 and who also voted YES
for some common subset of groups listed in 1c equals or exceeds 2/3
of the total number of people voting.
These three conditions aren't mutually exclusive; it's possible that one,
two, or all three could be satisfied. If (i) is satisfied, then it renders
(ii) and (iii) moot, obviously.
> > The only way that (a) can cascade into (b) without subjective
> > interpretation is if (a) comes out to be 2/3 or greater and renders
> > (b) moot, right?
>
> I disagree, please read on.
>
> > > If 60% say remove
> > > from lugnet.*, and 10% say remove from lugnet.cad.*, then you have a 70%
> > > super-majority for removing from lugnet.cad.*.
> >
> > That's one possible interpretation.
>
> That is my interpretation.
You miss my point. My point is that there *exists* some interpretation
which is external to the CFV document, not that any one interpretation is
holds more validity than another. In fact, there exists at least three
possible interpretations external to the CFV.
There shouldn't *be* interpretations external to the CFV document. It
should be 100% clear to the voters precisely how the votes will be tallied,
otherwise the process is unsound.
> Note when the 60% are saying YES to lugnet.*
> they are also saying YES to lugnet.cad.* and c.
It does not state that in the CFV. A voter has no way of knowing that
unless he/she guesses or assumes it to be the case.
> Also note that the 60% vote
> for lugnet.* would not have been a 2/3-majority and would have failed to
> pass, however, the 60% who voted for a + the 10% who voted only for b would
> result in a 2/3-majority and b would have passed.
Yes, I'm well aware of that; I'm the one who pointed it out. However, the
above is true only under one possible interpretation. If the vote tallier
wishes to cascade 1a into 1b & 1c, then the vote tallier should make that
100% clear to the voters, hopefully either through a revision to the CFV
document or as an addendum to it.
> > Another interpretation -- the literalist interpretation, which is how I
> > would definitely interpret it -- is that 60% want either complete removal
> > or no removal at all. That was the question as literally written.
>
> I'm a fairly literal person. Was the question literally written that way?
I'm not sure. That's an interpretation from one angle (which is bad -- the
point is that anything left to interpretation is bad).
What the CFV says is:
1. Would you personally enjoy the lugnet groups more if Jonathan Wilson's
posting privileges were revoked from:
a. All lugnet.* groups
b. Just lugnet.cad.* groups
c. Other (please list)
Ambiguity #1 (major): It doesn't state whether a, b, and c are mutually
exclusive or separate voting items.
Ambiguity #2 (medium): If a, b, and c -are- mutually exclusive, then it
doesn't state whether a vote for 1a also implies a vote for 1b & 1c
(cascading subsets) or a vote for 1a does not imply a vote for 1b & 1c
(independent all-or-nothing sets). In other words, it doesn't state whether
a=10% + b=60% equals b=70% or b=60%.
Ambiguity #3 (minor): It doesn't say whether lugnet.cad.* includes
lugnet.cad and its subgroups, or only subgroups of lugnet.cad (because of
the "." preceding the "*"). By convention, it would seem to, even though
literally (technically, that is) from a regular-expressions point of view,
it doesn't. A non-ambiguous form would be lugnet.cad* without the second
"." character, or better yet, "lugnet.cad and all of its subgroups."
> So far it doesn't seem like any of the voters interpreted the question the
> way you describe.
In other words, no one sent in a=YES, b=NO ?
> Which makes me believe the question was literally written
> for the other interpretation.
But _which_ other interpretation? Even without ambiguity #2, there are
still two interpretations subject to ambiguity #1.
> (By definition 1 in my dictionary Literal
> means "In accordance with, conforming to, or upholding the explicit or
> primary meaning of a word or the words of a text." According to this
> deffinition I would interpret lugnet.* as literally meaning every single
> group in the LUGNET newsgroups. Literally can mean "really; actually".)
> (Boy I'm literally being literal :)
I agree with that 100%; I don't thin there's any question whether lugnet.*
means every single group in the LUGNET newsgroups.
> However, if people feel that they want
> their vote to be an 'all or nothing' then please 'write it in'. This is one
> thing I should have mentioned in the CFV- 'write in's'. Just because the
> ballot has been set doesn't mean voters can't write in alternatives, like
> the 'all or nothing' interpretation described above.
In that case, it should explicitly say that if a voter doesn't write "all or
nothing," that it will not be taken as "all or nothing." If you intend
1(a,b,c) to be mutually exclusive, then it shouldn't be left up a voter to
wonder about that.
> But, keep in mind: if
> everyone writes in something without voting on the standardized questions
> then this vote will just turn into an opinion poll and none of the actions
> will pass.
> After reading Todd's post <http://www.lugnet.com/admin/general/?n=2569> I
> think a clearer way of writing question #1 to fit with my interpretation
> would have been:
>
> Qote from Todd's above post:
> " 1. Would you personally enjoy the lugnet groups more if Jonathan
> Wilson's posting privileges were revoked from (answer each
> separately):
>
> 1a. lugnet.* (as many as possible)
>
> 1b. lugnet.cad.* (as many as possible)
>
> 1c. as many as possible of a specific list of other groups
> (please list)
>
> Note: YES for 1a implies YES for 1b and 1c, even if you write NO
> for 1b and/or 1c or leave either blank."
>
> This is how I have been evaluating the votes, even though it is not stated
> like this in the CFV. In other words if you voted YES to a and NO to b and
> c, but you answered YES to 2 then I have been recording the vote as if you
> voted YES to b and c (to me your NO answers to b and c were just opinions-
> because you may not have been involved with those specific groups).
I think it would be good to post a follow-up to the original CFV message
noting this as a 1.1 revision, and allowing people to recast their vote on
the off-chance that anyone voted 1a=YES/1b=NO/1c=NO with the all-or-nothing
intent, as well as for anyone who thought that 1(a,b,c) might have been
mutually exclusive.
The thing is, the original CFV *is* ambiguous, so people who wondered
whether 1(a,b,c) were mutually exclusive probably hedged their bets and
filled in all three, figuring it certainly couldn't hurt.
> > As Larry pointed out,
> >
> > http://www.lugnet.com/cad/dev/?n=2672
> >
> > it would be "more straightforward to present a series of up/down
> > propositions and commentary that the broader ones will supercede the
> > narrower ones if passed."
>
> That's unessessary. If the broader one does pass then all the narrower ones
> will fall under that umbrella.
We'll be lucky if it's unnecessary. If the broader one does pass all by
itself, then it renders the others moot at the tail-end. But the broader
one certainly doesn't pass the narrower ones on the input-end. It's NOT
worded that way currently.
> > At this point, if people are considering a and b to be mutually exclusive,
> > I just hope it turns out either that a>2/3 or that a+b<2/3.
>
> b and c are only exclusive from a in the sense that some voters don't want
> all JW's privileges revoked. They only want his privileges revoked from
> those groups they have chosen. Also, for those who answer YES to anything
> in 1, but NO to 2, your votes DO NOT go towards those categories (in this
> case you are just expressing an opinion, your answer to 2 decides whether
> you want that opinion to count).
Great! :-) When the tally is announced, please include the above in the
summary -- as it's critical to the tallying function.
> I hope this has clarified some of the issues about this vote. I tried to
> make sure the questions were as clear as possible (and I did work on the
> questions with other people before posting, but I take full responsibility
> for the problems all this has caused).
A good portion of it falls on my shoulders as well, since I was the one who
suggested the wording to Adam which turned out to be ambiguous upon later
reflection.
> [...]
> Alternative:
> 3. Should Jonathan be placed on probation for a period of 6 months as an
> alternative to revoking posting privileges?
>
> If you answer YES to question 3, then please vote on the Violation of
> Probation definition.
> Violation of Probation is defined as:
> (please choose one of the following or write one in for d)
> a. 1 new post
> b. 5 new posts
> c. 10 new posts
> d. Write in
> by Jonathan that result(s) in a flame message being sent by someone in [...]
Questions: What is a "flame message" -- how is that defined? Who
interprets (judges) the definition against a given message?
Or are those details to be worked out later if and only if #3 passes?
--Todd
|
|
Message is in Reply To:
 | | CFV Clarifications and REVISIONS
|
| Hi everyone, I've finally noticed the concern about this growing so I'd like to throw out some clarifications: (I chose Todd's message to reply to because he seemed to address most of the concerns.) Todd Lehman <lehman@javanet.com> wrote in message (...) (25 years ago, 18-Aug-99, to lugnet.admin.general)
|
101 Messages in This Thread:
(Inline display suppressed due to large size. Click Dots below to view.)
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
