Re: [Gnu-arch-users] james on Canonical

[Thomas Lord]

 Tom> In what is at least graceless and arguably unprofessional ways
 Tom> y'all screwed both the users and the upstream project,
 Tom> thoroughly.

 Adrian> IMO, they didn't screw the users -- at least, not by
 Adrian> abandoning baz, which I have to assume is what you mean.
 Adrian> Unless they found some magic legal way to escape the GPL,
 Adrian> anyone is free to pick up their work and continue or extend
 Adrian> it, same as Canonical was free to pick up yours.

 Adrian> Also IMO, they only "screwed" tla by legitimately attracting
 Adrian> more development interest.  As far as I can see and
 Adrian> extrapolate from the posts on this list, they did this by
 Adrian> themselves paying people to work on it, by accepting patches,
 Adrian> and by creating a more appealing development platform.

 Adrian> Why is this bad?

Because there is more to a project than a static snapshot of its
source code and the dynamics is where most of the value is for all but
the most venerable/nearly-completely-stable components.

Canonical arrived on the scene confronted with what was oriented to be
a long term project.  That long term project was characterized by a
certain conservatism, in order to make sure it didn't get itself or
its users stuck up a tree.  The project had produced considerable
value, as evidenced not least by the application Canonical itself made
of it.  It hadn't yet found a real business model but at least it was
muddling through based on adoption number * busking-success-per-user
rate.

A lot of divergent user interests were being balanced and, yes, that
meant among other things that a certain level of frustration was
spread around.

Forgive me for being immodest but casually destroying that emerging
feedback cycle, with the brunt of that destruction falling on me, the 
guy who best understands Arch, is an awefully bad outcome.  If all
else being equal, it was an unecessary, forseeable, and preventable
outcome -- I don't think "legitimate" is the right adjective to
describe Canonical's behavior.

Canonical came in to do one thing I find quite legitimate, but
then in fact did other things I don't find quite legitimate.

It was quite legit to use arch and to hack it for their own needs.  It
was quite legit to cut corners in their own version where it was
consistent with their own needs.  It leaves legitimacy behind to
promote their corner cutting approach the ways they did only to hang
their users out to dry and, meanwhile, utterly disregard being
cooperative with upstream (which would be me).

I understand that personal feelings came into play there.  Confronted
with some criticism of their work, some of the Canonical staff took to
the response of vendetta.  Yet subsequent developments, including
Canonical's abandonment of their own fork and revc's demonstration of
a better path, show that the criticism was well grounded (at least).
The destruction was gratuitous and deserves the label "graceless and
arguably unprofessional".


 Adrian> If "Canonical" were merely the handle of a user, or a group
 Adrian> of users, I would praise him/her/them without reservation.
 Adrian> "He" took a project that had lots of contributors but little
 Adrian> overall progress (that I could see at the time), branched it,
 Adrian> brought in lots of contributions, and kept it going at full
 Adrian> tilt.  When "he" inevitably lost interest, it was up to
 Adrian> someone else to pick it up and keep going.
 
 Adrian> [...]

 Adrian> This happens all the time.  It's the blessing and the curse
 Adrian> of free software -- if you don't like how it works, change
 Adrian> it.  If you do a good job of satisfying your developers /
 Adrian> users, then people will migrate to your project.  If you
 Adrian> don't, someone else takes your place, whether you like it or
 Adrian> not.

 Adrian> Is it somehow suddenly different because "Canonical" is a
 Adrian> company?

I dissent from the view of "little overall progress".

Regardless, yes, it is different because "Canonical" is a company.  It
would similarly be different if Canonical were a non-company or 
non-profit effort operating similarly to Canonical.

The difference is relative power and how power is used.  Canonical
commanded a superior labor pool and therefore could overwhelm by
volume.

They faced a choice of deploying that labor pool with social
responsibility and with respect to upstream -- or doing what they
did instead.

An individual hacker, on the other hand (Walter Landry, anyone?),
has far less destructive potential.   The burdens of responsible
participation or dissent are correspondingly reduced.

 Adrian> Writing free software full-time and relying on donations and
 Adrian> sponsorship to keep going is an inherently risky and unstable
 Adrian> proposition.  The money may be enough to get by for a while,
 Adrian> and one might enjoy the work, but if donations dry up, one
 Adrian> could also be suddenly "out of a job" at any moment with no
 Adrian> warning.

It has never been my goal to rely on donations.   I think I've been
clear about that all along.   It has been necessary at various times
and enough to muddle through on.

The risk I took up front, on my own, was the first release of larch.

If the FOSS industry leaders were any good at evaluating new R&D and
targetting investment, that would have landed me an average (or 
slightly senior) ordinary hacking job and I'd have lived happily 
ever after.  They aren't and it didn't.  I was financially stranded
in a pretty nasty job market, and I was forced to keep doubling down
my bet.

Canonical's arrival proved two things:  yes, my R&D was value creating
and worthy of investment.  No, our FOSS industry investors haven't
learned how to not overgraze the commons.


 Adrian> I'm honestly considering studying today's approaches and
problems
 Adrian> (with arch as a primary basis), then making my own (on my own
 Adrian> time).  Naturally, I hope this does not further "screw"
 Adrian> arch. ;)

Just don't take your lessons about how to wield finanical "power
tools" from Canonical.  Respect people whose work you derive value
from, especially in those cases where free software licensing leaves
that as a choice you have to make on your own.

-t