[Gsoc] Two applications from same student with "likely to be accepted" status

[Martijn Faassen]

On 4/5/07, Aroldo Souza-Leite <asouzaleite at gmx.de> wrote:
> Lennart Regebro schrieb:
> > On 4/5/07, Baiju M <mbaiju at zeomega.com> wrote:
> >> Yes, at last one application will be chosen from a student, even if he
> >> submit  applications
> >> to different organization.  I think google is not going to look deep
> >> into our proposals, rather
> >> how we manage these applications is more important.  Google will only
> >> allow a particular number
> >> of slots for us.  I am afraid that the current proposal list will
> >> negatively affect us to loose one slot.
> >
> > OK, so we need to:
> > 1. Decide which of Pauls to psh up. It seems the most popular one is
> > the ZCML improvements. I put a -1 on that because I don't think ZCML
> > speed improvements is that important, but it seems that the speed
> > improvements would also mean that there would be some sort of python
> > interface for it, which I do find important. So if this can be
> > confirmed I'll add some plus votes on that one.
> >
> I like the Zope3/AJAX project better if the mentors would push it
> towards very general results for Zope3 AJAXability, not depending on one
> or the other special aspect. If Zope3 gets sexy, it is easier for me to
> find budgets to refactor applications now running (well) on Zope2 into
> Zope3.

That's a good point, and I think AJAX is important. The problem is
that I'm having some doubts about both actual proposals involving AJAX
(as mentioned in the google SoC interface too).

They seem to be focusing on one restricted use case involving deep
integration into Zope 3's formlib.  A use case that I'm not sure I
have that much use for, though a better motivation may make me more
positive.  I can certainly see
lots of use cases of AJAX that this use case doesn't cover.

How does all of this relate to KSS? Why no effort to create rich AJAX
enabled widgets instead of enhancing formlib this way? The gain in
pretty dynamic user interfaces seems to be much bigger gain for the
effort if those routes were explored.

Finally, Formlib is not the be-all and end-all of form libraries and
the investment may not be worth the effort. Perhaps I'm
misunderstanding what both proposals actually imply, but right now I'm
worried.

The ZCML proposal on the other hand does mention a Python API, which I
would like. Getting a speed increase in startup time for Zope 3
wouldn't be bad either.

Regards,

Martijn