[Grok-dev] Day #1 of the Grok Sprint

[Brandon Craig Rhodes]

We enjoyed seeing Lennart Regebro and Stefan Richter when each of them
dropped by our sprinting room here at PyCon yesterday!  We are sharing
a room with the Enso folks, as well as with the Pyglet team - which
brought about twice as many people as we did!

Django is definitely the PyCon web framework of choice, and from what
I have seen is holding the largest sprint at the entire convention;
they seem to have around forty people participating.  Turbogears lags
behind them, with a contingent of maybe between two and three dozen -
large enough, though, that they had to move from their assigned one
into a larger one.  We Zope/Grok folks, on the other hand, reached a
peak participation level of five people.  If I were new to Python and
deciding on a web framework here at PyCon, it certainly wouldn't
appear as though Zope were worth my time.

I suppose that it's because Plone has its own conference, that people
would rather attend, that it's completely invisible here at PyCon? :-)

Anyway, yesterday I and Robert Marianski created "grokcore.component"
and got it mostly whittled down to a working subset of Grok.  We have
not yet tackled what on earth do to with the interfaces.py file (we
suppose that some sort of inheritance hierarchy will now have to exist
between the various grokcore interfaces and the main Grok interfaces),
and it's this morning that we'll actually rename the module now that
the test cases are working with the additional functionality, and then
try using it against a separate project and create a howto document on
using the component.  (We ought to add a "Grok Components" section to
the documentation site, where we can place guides to using various
pieces of Grok by themselves.)

Brad Allen, Reed O'Brien, and Peter Bengtsson all worked on
documentation, doing excellent work reviewing and improving the main
Grok tutorial.  They also asked some interesting questions - about,
for example, whether one can really move from Grok into taking full
advantage of Zope 3 without rewriting an application.  I think there
might be a need for a long tutorial document that takes a basic Grok
app, then shows how one would pull in the most advanced features of
Zope 3 without having to rewrite it.

I'm hoping to finish up my grokcore.component work today (I'll
probably need help releasing it as an egg if I make it that far), and
after looking at the very high number of Django people, I'm seeing
that people who are generally not like me - I chose Grok because the
underlying framework uses, in my opinion, the most powerful yet
tractable OO abstractions (adapters, etc).  But most people want to
make pretty and useful web apps for customers quickly.  And right now,
our tutorial doesn't show anything pretty.  I wonder how hard it would
be for the tutorial to get the student very quickly to a small web app
with, say, a date field with a little pull-down calendar, or a text
selection box that pops up autocompletions dynamically?

-- 
Brandon Craig Rhodes   brandon at rhodesmill.org   http://rhodesmill.org/brandon