Sam Ruby wrote:
Sylvain Wallez wrote:
Adam Peller wrote:
AJAX Toolkit Framework Proposal
I'm quite puzzled by this proposal. As I understand it, its mainly
about a set of Eclipse plugins for Ajax applications and the Zimbra
library that, among other features, provides a set of SWT-like widgets.
Yes.
Also, this proposal pops up right after I mention on members@ that
several projects at Apache are using or plan to use Dojo [1] and that
we talked about inviting them. I sincerely hope this is just a
coincidence.
Completely a coincidence. I've been aware of the plan to submit this
proposal for several weeks, and hadn't seen your post until you
mentioned it. I also had a conflict that precluded me from coming to
the ApacheCon.
As a general rule, the ASF doesn't go out "inviting", people within
the ASF either start a new project, or projects come to us.
You're playing with words. Sure, there's no formal invitation process.
Now ASF members can approach projects they find interesting and "suggest
them to submit a proposal to the ASF", for the greatest benefit of both
the coming and existing ASF projects.
Thinking more about it, the fact that the ASF isn't supposed to invite
projects seems to go against the ASF meritocratic rules. You should not
ask for being a committer: you are voted in when other committers
consider you deserve it. And you can reject the offer. Same for
membership. Why couldn't it also apply to projects that already follow
the Apache way and are of interest to the Foundation's projects?
On the other hand, proposals like this one, originating from commercial
entities, really look to me as "pushing the ASF door open", even if the
incubator is supposed to ensure community diversity and healthiness
before graduating as a real project.
In any case, the ASF is not exclusionary: if there was interest Dojo
could be added to this proposal, or could pursue a separate proposal.
Right. Now I don't consider starting a proposal war to be the best thing
to do. Especially considering that one of the Dojo devs told me "Those
[the ASF benefits] are all good things, however the political and
organizational overhead of the ASF appears huge". Bingo.
So the questions are:
- is the ASF the place for Eclipse extensions? I don't deny the
ability to _existing_ project to host their tooling, but this isn't
the case here.
As I mentioned, I was involved with these discussions. The ASF
doesn't tend to make these types of decisions based on the technical
aspects of a project. What impressed me about the people who were
proposing this is that they were sincerely interested in the Apache
License and collaboration model.
While the Eclipse development model is certain a valid one, it is
different in a number of significant ways from the ASF. Suffice it to
say that I am partial to the way the ASF does business.
Ok. Now some of the planned features seems to directly overlap with
what's already in webtools (e.g. the JavaScript editor), and this
project would be the first one at the ASF in the general IDE tooling
category, which is what Eclipse is all about.
Sure, the development models are different and Apache cares about
community and not technical details, but this seems weird anyway and I'm
wondering if that won't turn into an OSS organizations war which would
certainly be detrimental to all of us.
In other words: why isn't this IBM-originated generic Eclipse tooling
donated to the Webtools project, that also originated from IBM?
- why incubate an Ajax library that none of the current ASF projects
uses nor plans to use, unless I missed something?
It is a valid question, but it is also valid to point out that the ASF
has projects as diverse as TCL and SpamAssassin.
The situation is very different here: several projects are integrating
Ajax features and incidentally found that they were considering the same
framework for that purpoe. Whereas none of the ASF projects was already
envisioning close integration with a spam filter when SpamAssassin came
to Apache.
That could even end up with the funny (ahem) situation where Apache has
an Ajax framework that isn't used by its Ajax-enabled server-side
frameworks. Doesn't it sound weird?
What is more important is considerations that the code be licensed
with the Apache Software License (not dual licensed, like Dojo), that
the committer bases be diverse, and operate in an open and
collaborative model.
C'mon! The incubation process is meant to solve licence and IP problems.
Zimbra is MPL & ZPL(?), and the IBM contribution is "Licensed Materials
- Property of IBM"!!
The Dojo peeps dual-licensed their stuff to allow the widest
distribution possible [1], and have a development model very close to
the Apache way, with active user and developer lists, and committers
nominated on a meritocratic basis. I can't see the same in this
proposal. The Zimbra stuff, as technically impressive as it can be, is
the creation of a single company whose commercial offering is based on
it. Nothing that prevents it to incubate of course, but community
diversity isn't an easy thing to achieve in such conditions.
Sylvain
[1]
http://blog.dojotoolkit.org/2005/12/04/dojo-now-dual-licensed-afl-and-bsd
--
Sylvain Wallez Anyware Technologies
http://bluxte.net http://www.anyware-tech.com
Apache Software Foundation Member Research & Technology Director
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
