[KLUG Advocacy] Re: [KLUG Members] XFree86

[Peter Buxton]

--opJtzjQTFsWo+cga
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable


To advocacy:

On Fri, Mar 21, 2003 at 10:30:16AM -0500, Adam Williams wrote:

> I did, and several related articles and mail threads (including
> Keith's rebuttal).
>=20
> My conclusion - somebody should slap Mr. Packard and tell him to get
> with the program.  This is just like what happened with Mr.
> Kennsington and the Samba Team,  and look how far Samba-TNG has gone -
> no where.

Not true, I'll bet. I'm sure whatever CIFS codes they learned were
shared with Samba.org.

Also, Andrew Tridgell's wise and mature blessing of samba-TNG is
light-years removed from XFree86 core's Gallic pronunciamento: "I hate
people who do their own X development! What, you're doing your own X
development?! And you didn't tell me! You're stripped of commit-bits!"
Pathetic. And even more:

http://www.xfree86.org/pipermail/forum/2003-March/000268.html :

> > but there are quite a few "Stake-holders".  I happen to be one of
> > them, and I replied to your request for input as a concerned
> > stakeholder.
>=20
> I am glad to hear that!  I presume that "stake-holders" like yourself
> will be willing to put your money where your mouth is in exchange for
> having a say in how XFree86 is run.  Maybe as a sign of good faith you
> will be giving me time to work on XFree86 and its releases?  That would
> do a *lot* to help alleviate the resource problems release cycle issues
> that have been identified.  Like I said in a previous message, when I
> worked for VA, releases came out on a regular schedule.  Since then,
> with Tungsten, they have not.  Holding a stake cuts both ways.

In other words, "Pay me or you don't get your X the way you like it."
Not politic, to mention bribery so publicly, but hey -- tres Gallic. (I
wonder if Mr Dawes ever filled out one of those, whattayacall, job
applications and sent it in to Mandrake or Red Hat.) Or maybe RH (or HP,
Packard's employer) has their own employee they prefer to subsidize
XFree86 support with.

No, please, don't compare Samba-TNG with XFree86 core's rude dismissal.
Mr Tridgell is in a class far superior to core. He maintained good
ties with Samba-TNG and kept his team on track. (Do you think Linus is
threatened by RTLinux? By the -ac tree?)

> Thank the PTBs that ATI can't commit code.  Xfree isn't something you
> just want people mucking about in, or using for their own agendas.  It
> is too critical and crosses too many platforms for your average
> yahoo-pc-jockey to be hacking on.

I agree. ATI/3dFX and nVidia should not be allowed to write their own
drivers. Neither should they be allowed to put in a driver side-by-side
with XFree86's driver for the same cards.  And, of course, they would
abuse that commit bit to write garbage into XFree86 source files other
than their own drivers, wouldn't they, the little chip-monkeys?  They
are, after all, only pissant yahoo-pc-jockeys.

So you'll be switching to OpenBSD, then? Linus entrusts entire bits of
the kernel to third parties, some of them hired by irresponsible
companies not unlike ATI, and I hear on the list he only wants this
behavior to increase. Tsk, tsk. Theo wouldn't allow that kind of
promiscuous code-sharing to go on.

> The comparison to GNOME and KDE is spurious.  XFree does *1* thing,
> provide a display architecture.

And not very well. It is clearly the most visible, and most primitive,
piece of the OS/GUI sum (I speak of the kind of sum that is greater than
its parts, analagous to LAMP). Mr Packard has completed several projects
for X with excellent results, so he is unlike the ill-fated Rik van
Riel, who screwed up his first major Linux project and nearly failed to
live it down.

Think of the number of extra projects designed to try and replace X.
GGI, Fresco/Berlin, SDL. If XFree86.org were healthier, these would be
centered around it or actively contributing to it.

Get rid of this idea of video as black magic. CHILDREN program neat
graphic demos 24/7. Sure, they aren't all the next John Carmack, but I
imagine some of them could further improve X.

> > Word has it that Keith Packard does *not* wish to fork XFree86, as
> > core accused him of doing. But he does seem to have talked to some
> > groups with a vital interest in X on free *nix (KDE, Gnome, GGI,
> > SDL, and the
>=20
> He did this without consulting his fellow team members, they had no
> idea what he was doing - that is unacceptable.

I agree. Obediance is a virtue (everywhere but France).

Until it isn't. I have to admit, I've really enjoyed skimming lkml. I
admire Linus Torvalds' style. (Ruthlessness, when necessary, is a
virtue.) But the thing I most like about his style is that it as totally
public as lkml itself. Nothing hidden, balls out. Brilliant.

XFree86.org will either become more responsive and open, or it will be
replaced by a group putting out better code faster. I think Keith
Packard will be there either way, core notwithstanding.

> > Linux distributions, I *imagine*) about creating an XFree86 adjunct
> > to speed up the coding. Alan Cox has weighed in on Packard's side:
>=20
> He agreed with some of Keith's concerns, and they may be legitimate.
> I think it is a little strong to say he took Keith's side.

Of course. He just showed up on the mailing list seconds after it was
created by "coincidence."

Havoc Pennington,
http://www.xfree86.org/pipermail/forum/2003-March/000240.html :

> It's more that the desktop/toolkit projects have assumed that X.org is
> hopeless and XFree86 isn't interested. Either of those assumptions may
> be wrong.

We'll see.

--=20
-182
Power tools for power fools.

--opJtzjQTFsWo+cga
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQE+e90zaAK8ZwqLoskRAhU4AKCayDlUlZA0FyRh7oRHENS1qYR77wCfSJJ/
KOVQRcgJ50Ry5r1Oy7tcEmI=
=16J7
-----END PGP SIGNATURE-----

--opJtzjQTFsWo+cga--