What's Canonical thinking about Bazaar?

[Martin Pool]

2009/11/4 Ben Finney <ben+bazaar at benfinney.id.au>:
>
> What disadvantages Bazaar is the perception that, instead of “Bazaar is
> a community project (and thank you to Canonical for continuing to make
> that possible)”, rather we have “Bazaar is a Canonical product (and the
> community is along for the ride)”.

Really, how much people are along for the ride is up to them.  If
people put up patches or designs we will engage with them and they can
influence the project that way.  (We don't always review them as fast
as we would wish, and we would like testing to be easier, but we're
working on both those things.)

There is no limit on how much someone can contribute without being a
Canonical developer.  One of the coolest and most powerful features of
Bazaar, the foreign branch plugins, came from a non-Canonical
developer, Jelmer.  (We've later helped support his work.)

People can have a useful contribution through only opinions but only
up to a point.

> Note that *it does not matter* to what extent this perception is true;
> it is the *perception* that is harmful.
>
> This disadvantages Bazaar in comparison to technically-comparable VCSen
> like Mercurial and Git precisely to the extent that the perception
> remains that those other VCSen are community projects but Canonical has
> a controlling interest in Bazaar.
>
> Please, please, if you will do nothing to quell that impression, at
> least do nothing to further entrench it. The Bazaar community has little
> influence over this matter as it is; we don't need to be fighting
> against Canonical marketing on the issue.

I'd like to pick apart the problems here because I think "Bazaar
suffers from being too closely tied to Canonical" is too vague to
either rebut or address or change.  Merely steering straight away from
it is unlikely to help.

If it's something specific like hypothetically "non-Canonical
developers don't get useful timely reviews" or "you don't care enough
about Windows" or "you're talking too much off the public list" or
"I'm afraid bzr 3.0 will be proprietary" it's something we can work
with.  It may be nontrivial to fix but it seems like a more fruitful
conversation.

>>  > I also believe that "lock-in" fears are largely unfounded.
>
> Again, nobody in this discussion has raised that spectre to my
> knowledge. Lock-in isn't the problem; perceived lack of community
> ownership is the problem.
>
>> And if getting contributions from non-Canonical sources is a *goal*
>> (vs. merely "actively encouraged"), I think the already high
>> concentration of development activity in the Canonical group is going
>> to work against achieving that goal.
>
> Note that, while a detrimental way this could be addressed is to reduce
> Canonical develpment involvement in Bazaar, a far more positive way
> would be to actively encourage the perception (which implies actively
> discouraging contradictions) that Bazaar is owned, not by Canonical
> corporation, but by the community.

"Owned by the community" sounds nice (though somewhat socialist) but
what does it actually mean?  (Not a rhetorical question.)

For instance, how do you think we should decide what features to do
next?  I've never heard of an open source project where objectives are
set by some kind of domination of the noisiest users over the
developers, and I don't think I'd like to join one.  Obviously you
should listen to users, but ultimately the developers will work on
what they think is interesting/important, or what someone pays them to
work on, or ideally they are the same thing.

If a core developer decided they cared, purely on a personal whim, to
concentrate on features that would help Ubuntu, would you tell them
they're wrong?  Where would this stop?  "I never have memory problems,
don't waste time on that!" "How dare you not fix my bug?"

We could potentially have a steering committee or a foundation, but
we'd need to have a view of what they would actually do differently,
not merely that there should be a box with that label.

I'm pretty skeptical about generalities like "community ownership" but
aware that we can do the specifics much better.

-- 
Martin <http://launchpad.net/~mbp/>