On Jan 25, 2006, at 12:24 AM, Raphaël Luta wrote:
As far I can tell, what needs to happen to fully OSS AjaxTk
is something like this:
- cleanup the code/doc/install so that Tk can be consumed
by a public community
It seems to me that something that works in some cases but isn't fully developed as an end-user easy-to-use product is ideal to complete in the incubator.
- rework a clear internal separation between AjaxTk and
ZCS with different repos, dependency management, indep
release cycle, and so on...
Repository issues, build, test, separation of dependencies, etc. can be very repository-specific. If you build this infrastructure in source forge you would have to deal with similar but different issues here in Apache. Why require them to make two transitions?
- adjust internal dev processes to deal with the public
infra (and possibly some early external dev)
IMHO this is a critical part of building a community.
- start building a long term external community by
actively attracting new committers
I would feel more comfortable if the first 3 steps would be
done somewhere else before any incubation starts so that
incubation can really focus on the last point.
Sorry, but I don't agree. The way the first 3 steps are handled will give the Apache community a good sense of whether this project "gets it" or not, and we might as well help them with the process sooner than later.
Apache does not bring any value in the first 3 steps that you
can't bring by yourself as an ASF member and the amount of
work and risk of failure in those 3 first steps is IMO
significant.
I'm not sure that I understand your point. Certainly "any" open source repository is a different environment from a closed source, but I don't see an argument for or against incubation in Apache.
The benefit of such a 2 step process is that this gives
you an actual public track record before incubation is
decided and voids most concerns about incubation being a pure
branding exercicse since a significant amount of work would
already have been spent in making the AjaxTk OSS independantly
of any Apache commitment.
I'd still say that there is significant breakage involved in taking a closed source through two migrations rather than one. I think we need to remain focused on the barrier to incubation as documented in the Apache policies and not try to create new processes on the fly.
To me the critical piece here is "donation". These folks have intellectual property that appears to be useful and they want to donate it to Apache. Due diligence is required to make sure that there are people in Apache who want to see it succeed (mentors, sponsors, champions, developers, users), and there are resources available to build a community here. But let's not put artificial barriers in their way. There are plenty of barriers once in incubation.
smime.p7s