Re: [Gnu-arch-users] Arch funding models, again

[Matthew Palmer]

[A long-brewing diatribe follows.  Enter at your own risk.  Sorry for
attaching this to your message, Robin, but your message just turns out to be
the trigger.]

On Sat, Sep 25, 2004 at 12:00:33AM +0100, Robin Green wrote:
> Obviously there are other problems here, but my point was, we don't appear
> to have a reliable gnu arch maintainer any more, which is what some users
> would like to see. _One_ way of fixing that - in fantasy land! ;) - would be
> to pay someone to do it full-time. (Of course paying people doesn't

It would appear that there will always be problems getting anyone other than
Tom to do the complete release process for tla, which means that we can pay
someone to do what jblack has been doing (merging, testing, RCs, and so on)
but actually getting a Real Release of tla out the door requires Tom's
say-so, which (as I can entirely empathise with) he doesn't want to give
without thoroughly vetting the changes he is putting his name to.

That last little bit there is the real problem, in my humble and incomplete
understanding of the situation.  tla really is "Tom Lord's Arch" -- and Tom
doesn't want anything out there with his name on it that doesn't meet his
high personal standards.

So, my proposal (impractical though it may be) would be to "fork" tla in a
way, basically just renaming it to something like taca -- "The Arch
Community's Arch".  I'm firmly convinced that, like the Linux kernel is
always associated with Linus, Arch will always be firmly associated with
Tom, but from what I've seen, Tom is a blocker for smooth, small "early and
often" releases.  Sorry Tom, but everything I've seen on this list and
elsewhere about Arch leads me to believe that you're a "messiah type" --
lots of really cool ideas that will make the world a better place in the
future, but not quite so much of the housekeeping tasks that keep the world
spinning in the meantime.

Please don't take that as a criticism, Tom -- the world needs visionaries
and "uber hackers" to take things in new and weird directions to the benefit
of the wider world (which, in the revision control sphere, I think Arch has
done).  But the housekeeping things like regular releases and keeping
patches merged and the million other little things need to be done all the
time.  Releases can't stall for a few months while you go and work out how
pika is going to change the world.  It appears that Arch has become too big
to be happy with sporadic releases and a few militant hackers, all with
their own compiled versions, cherry-picking each others' patches as the need
arises.  The Greater Arch Community needs something to hang their hat on,
and IMO that is a regular, stable, featureful release of an Arch
implementation.  I'd call it tla, but it *shouldn't* be simply TL's A, it
should be *everyone's* A.

Funniest thing is, I'm pretty happy with the current tla feature set -- I'm
running 1.2-4 from Debian and am quite happy with it.  So a complete lack of
future releases wouldn't actually harm me particularly.  I wonder what I'm
getting so worked up about...

- Matt

signature.asc
Description: Digital signature