[Compcomm] Deciding a name

Sun Jun 10 20:16:22 EDT 2007

Am Sonntag, den 10.06.2007, 23:55 +0100 schrieb Nicholas Thomas:
> On Sunday 10 June 2007 21:19, Kristian Lyngstøl wrote:
> > On 6/10/07, Nicholas Thomas <nick at lupine.me.uk> wrote:
> > > On Sunday 10 June 2007 19:26, Kristian Lyngstøl wrote:
> > > > Many people are not satisfied with how we're voting on a name, we
> > > > should respect that.
> > > >
> > > > I therefor suggest we void the current vote, and restart the process.
> > > > To make it fair and democratic, this is how I suggest we do it:
> > >>
> > >> ( .....)
> > >>
> > > > Input on this?
> > >
> > > Sounds like not-a-plan, because we'll likely end up with no name
> > > whatsoever.
> >
> > Then we need to figure out a new name.
> >
> > > With the current poll, just under 1/4 of voters aren't happy with the
> > > choice of names available (i.e. voted "none of these"). And, you know,
> > > that's not particularly great, but it's not particularly bad, either.
> >
> > Out of all the voters, I don't care what most of them say, because
> > most of them are NOT involved with this project except as users.
> >
> > Rico and Mike have both contributed with actual work, I care more
> > about their opinion here than random users.
> >
> > > To keep it short - the arguments against coral as a name are spurious, as
> > > has been discussed in great detail both here and elsewhere. It's a name,
> > > it has significant community support, and I think that if the poll ends
> > > with it in the lead, we should respect that and go ahead with the rename.
> > > 9 hours vs. 19 days... come on.
> >
> > If it takes 19 days to reach a decision everyone who has actually
> > ALREADY contributed to the project can be reasonably satisfied with,
> > then that's a small price to pay. That this wasn't done from the start
> > is not an excuse.
> >
> > The poll has put all the "filtering" responsibility on one person, who
> > hasn't had any set rules helping him. That's not fair for him, or the
> > community he's working for.
> >
> > As you should know, I do not care too much about the actual name, but
> > I do care about the process of picking it. And this vote has been
> > horribly random, not planned and not agreed upon before it was set up.
> > In short: it is a forced decision by a minority of the project that
> > many people are willing to go along with just to get it over with. It
> > shouldn't be that way.
> >
> > --
> > Regards,
> > Kristian
> 
> Moving to the ML, as requested. IRC log attached to a seperate mail, coming 
> through whenever it's approved by a mod.
> 
> My argument against voiding the poll is relatively simple. It's not a perfect
> poll, and effort could have been made to engage greater segments of the
> community, it's true -- for this reason, I'm OK with the idea of extending
> the poll to allow a greater proportion of the community to have their say..
> But I am against voiding it. Hold on to your hats, now...
Nobody said we should void the poll. We should re-start it with choices
that are less stupid/boring and more unique/creative, even if it takes a
month to gather some good name suggestions. There is absolutely no need
to hurry with a name. Call your packages "compcomm-*" or just "compiz-*"
- nobody cares about what the packages are named like.

> 
> Pre-merge, it was agreed that the name would be chosen by a democratic vote.
> That means that if davidr wanted to weigh in, or Quinn_Storm, they'd have as
> much say as SmSpillaz. One person, one vote.
Where and by whom was that agreed? Why do agreements you (whoever that
is in this case) made "pre-merge" apply to the new project? Even though
it was mentioned in one of the drafts describing "compiz-extra", I
wouldn't say that it has been agreed on.

> 
> We had a name suggestions thread with 20 pages of community discussion. Some
> people (notably from the old 'compiz' camp) weren't involved in this, it
> seems. We then had a name choice poll with the most popular of the names
> given in the name discussion thread, + the "none" option, which required an
> absolute majority in order for it to win. In the event that "none" won the
> vote, a second poll with a new set of names was to be set up. Again,
> engagement on this particular area was less than satisfactory.
I (as the only one from the compiz-camp as it seems) didn't take part in
the name poll because I think it is a big mistake (reasons summed up in
my previous mail). Before the poll was started I suggested to keep out
the combo-names - without success. As you can see they will most likely
win because of the big amount of (possibly younger) beryl users taking
part in the poll (coral sounds pretty much like beryl so they'll love it
anyway).

> 
> Now it gets a little philosophical. Bear with me. Every possible name was
> catered for, because of the inclusion of the "none" option. Therefore, any
> lack of engagement prior to name choice would not have affected the poll.
> Therefore, the poll is valid.
> 
> Why is this so? Let's take a simple example. 10 possible names; 1,000 votes,
> distributed evenly. 5 choices + "none" in the poll. "none" gets 50% of the
> vote (technically, this would be a loss of none, since it's not a majority,
> but that's a sheer technicality and we can all agree that in the event
> that "none" got exactly 50% of the vote, it would be taken as winning). So
> now we have 5 names left in our theoretical namespace. Same 1,000 votes; each
> name has a doubled number of votes. Obviously, it's a tie.
> 
> The important things to note from this example: Every name, in the absence of
> any other factor, got an equal number of votes, in the end. Which is to say,
> the probability of a given name being selected, disregarding factors external
> to the vote, is equal.
> 
> Moving back to our real example, we have a theoretically unlimited (but
>  finite in practice) number of possible names, and an unspecified number of
>  votes. Vote choice is determined by a wide variety of different factors. The
>  ONLY factor that we can account for is voter preference - that is to say, we
>  can assume that a user voted for a particular name (or "none", which is, in
>  effect, the same as voting once for every other name in the pool) because
>  they wanted that name to win. We can assume that because that's what the
>  poll told users to do.
> 
> Everyone still with me? Good :)
> 
> People now want to void the poll. They argue it is invalid because a
> proportion of the users weren't involved in choosing the name, so their views
> weren't represented in the poll. I argue that because every possible name has
> been included in the poll, the poll is valid, and so shouldn't be voided
> (note that I'm not arguing that the selection method is perfect, but that it
> is valid). Again, I have no arguments with giving the poll more time to run,
> although I think that more than 24-48 hours would be excessive.
> 
> Important note: Invalid arguments against this line of reasoning are any that
> include "but the users would have been affected by X because Y". Without
> asking - and receiving responses from - a representative sample of the
> voters, we can't make any such argument. The only reason I can make the
> assumption that voters have voted for their preferred option is because
> that's what the poll told them to do. We know that users are affected by all
> sorts of things, but without any evidence to back up any argument that
> involves second-guessing users' motives, there's no point delving into it..
> 
> But at the end of the day, you guys do whatever you think you should. If you
> don't believe the poll is valid, we can talk about it and maybe agree
> (although I think the above argument is pretty iron-cast). If you believe the
> poll is valid and want to void it anyway... well. Then we've got nothing to
> discuss.
> 
Sorry I don't see anything iron-cast in your argumentation. Nobody said
that "it is invalid because a proportion of the users weren't involved
in choosing the name, so their views weren't represented in the poll".

It is all about style. I think you don't understand how important a
stylish name and a well-designed project identity will be for the
project. Coral is an inacceptable solution and should never have made it
into the poll - this was a big mistake on our (the management's) side.
In case you haven't noticed - we all are (i.e. feel) responsible for
this project. The project is not only a piece of code - it should be
seen as a piece of art, in all of its tiny parts ... Would you coders
accept bad-styled or buggy code written by some newbies only to be able
to release the packages a day earlier? I hope you wouldn't. The same
applies for me and the name, I can't just accept a silly name if there
could likely be a much better one after some thinking.

We should avoid all those typical mistakes from other OSS-projects (e.g.
ugly website, ugly logo, stupid names) - look at blender.org if you want
to have an orientation how I'd define a (more or less) professional
project identity and a very well designed website and logo. 

I understand that a project will slowly evolve with the time but we have
the chance to save three years of "evolution" by simply thinking and
talking a bit before we actually start doing things. It worked (and
still works) well for compiz so why shouldn't it work for the new
project as well?

Democracy needs guidelines, otherwise it is pure anarchy ;) ...

Rico

> /Nick
> _______________________________________________
> CompComm mailing list
> CompComm at Rock3d.org
> http://www.ubaight.com/mailman/listinfo/compcomm