XFree86 teleconference minutes posted [LWN.net]

quote context

Posted Apr 4, 2003 17:53 UTC (Fri) by mattdm (subscriber, #18) [Link]

You should include more of David Dawes's quote. Out of context like that, it sounds like a nasty thing to say. Clicking on the link, however, reveals:

I don't know why you're all wasting your time talking about this stuff when there's nothing stopping you from just going ahead and creating your own project. I'm looking forward to it, and I'll be amongst the first to apply for commit access :-).
Haven't we all wasted far too much time on these fruitless "I said, he said" arguments that never have a resolution?
Speaking for myself.
David

quote context

Posted Apr 4, 2003 19:59 UTC (Fri) by rfunk (subscriber, #4054) [Link]

Yeah, David's full quote sounds better than just the first sentence of it, but taken in context of the posted transcript, it sounds like he's unwilling to make the changes that people want in order to AVOID a fork.

Or maybe he just missed the part about giving xfree86 a month to improve things and only then forking.

quote context

Posted Apr 6, 2003 15:25 UTC (Sun) by IkeTo (subscriber, #2122) [Link]

> David's full quote sounds better than just the first
> sentence of it

My English must be too bad to sense it. To me the two are equivalent: they are unwilling to give up governance.

Starting up another project is possible anyway. Starting another Linux is possible as well, just like starting another PostgreSQL, or Apache, etc. But anyone who witness how "another Emacs" is started knows perfectly that it is something really bitter. You mean spliting up the user base, having every feature implemented twice, every bug fixed twice, etc, not mentioning having strong religious wars between the two camps. I see "wasting" half a year, or even a couple of years, on such issues a better solution than wasting the years after that when we have every driver implemented twice, with some application available to one X implementation but not the other.

quote context

Posted Apr 10, 2003 15:23 UTC (Thu) by metacircles (guest, #8895) [Link]

A fork doesn't have to be a bad thing. Many people are fond of looking at the results of a fork and saying "oh, if only they'd worked together they'd have beenn able to do twice as much", but this doesn't admit that sometimes the reality is that if they'd only have stuck together they wouldn't have done anything at all. Forks arise where there's a conflict of goals, and if that conflict can't be resolved in a fashion which is a win for both parties, there is no shame and much advantage in letting one or other party branch off and do his own thing.

It helps if people are mature about it, of course - it doesn't have to be bitter and acrimonious. I'm involved with one project (CMUCL) which forked to form SBCL three years ago: now CMUCL has twice as many committers as it did before, targets another two architectures (it's a compiler) and has much more participation on its mailing lists. SBCL has four or five committers, none of whom were CMUCL developers anyway. And we watch each other's mailing lists and discuss the best way to fix bugs - which has the pleasant side effect that much more peer review of code gets done - and new features in one system are borrowed to fit into the other when the leaders of each project thinks they're good ideas. It doesn't have to be any more divisive than multiple Linux distributions are.

Given the strong interfaces between different parts of X (libX11 is documented, the X protocol is documented, even the driver interface has to be stable enough for binary-only drivers) I don't think a fork is going to impact users to the extent they have to choosing between supporting (as in, writing code that depends on) one or the other project, so the Emacs/XEmacs split (where the elisp interface is different between them) is not the obvious comparison.

XFree86 teleconference minutes posted

Posted Apr 4, 2003 19:39 UTC (Fri) by donbarry (guest, #10485) [Link]

It's increasingly clear that this fracture has been primarily along the
lines of an old-guard with attachments for reason of pride or reputation
and new and active contributors who primarily are seeking efficient
processes to integrate potentially disruptive but featureful
extensions. Wexelblat, who appears with Dawes to be the public side
of the group which purged Packard, contributes now only politics (he
admits leaving the working core of the project years ago, and now only
personally uses Windows platforms). The disruptive potential of
people like this is, I believe, one of the primary issues requiring
resolution [Would it satisfy the Core to save face by making Wexelblat
the sacrificial victim rather than Packard? :-) ]

Of course, there are exceptions. Eich joined core within the last five
years and seems to be a person who abhors politics, and therefore
naturally has a certain attachment to the status quo. Gettys' wisdom is
respected by many, and as perhaps the only current board member who
is both a greybeard and an advocate for a more inclusive project, I
believe he has a key role to play in either bridging this break with
real change, or if a fork is forced by the old guard, being the one who
in making the jump convinces others that whatever effort is joined
by Packard, this is the one for the community to sign up with.

Otherwise, it's clear to this outsider -- on the one hand is talk of
both technical and political issues which are entirely sensible for
the governance and architecture of X, and on the other is innuendo,
veiled character aspersions, and the ridiculous demeanor of those
who must cloak in more bitter forms what at most is only lese majeste
from those who have replaced them.