Lennart Regebro wrote:
On 3/1/06, Chris Withers <[EMAIL PROTECTED]> wrote:
Does this help with implementing persistent schemas at all?
Maybe, although I feel that the interface-based schemas was a mistake
anyway. It would probably be better to focus energy on making XForms
based schemas, both persistent and not.
Just my 0.02€
I'll pitch in my two centimes as well. Not only that, but I'll actually
work on the implementation. :^) I'm already doing XForms-style TTW dev
in a pet project.
I'm personally interested in TTW approaches that look at the problem
from a different perspective. I realize that (a) not everybody will
have that perspective and (b) trying to prod others to do it isn't fair,
that I should do it myself.
In a nutshell, my perspective is that the TTW environment should use TTW
technologies. It shouldn't try to make some semi-restricted, safe
subset of the on-disk technologies work correctly "TTW".
I also endorse the idea others have made of not making this part of the
core. Perhaps not even use the name Zope for the TTW package, similar
to how Rotterdam originally intended to have its own mini-brand.
In fact, perhaps the effort could be done without really being tied to
Zope at all. TTW dev that provided REST-style introspection and
updating. Glue code (I'm envisioning lxml's namespace extensions) that
make the object model and TTW model seamless.
Anyway, that's all just my silly little perspective. :^) Until I
actually show anything, it's moot.
Zope3-dev mailing list