Hi Jeff. Your approach is very interesting. There is lots room in Zope
for differences in the way applications are tackled to deliver on
requirements. My feeling is that explicit object metadata and some of
the content-management-esque concepts in Zope fit prominently into the
future of the web.
Zope's interfaces and adapters facilitate benefits that can be derived
from hybridization with any form of storage. As much as the ZMI may
hinder, I believe it plays a role in visualizing incremental development
and utilizing what is already there. Unfortunately, time and budgets are
also business considerations and this same infrastructure allows for
some fairly quick development.
It sounds that the ZMI and browser views hasn't been an impediment but a
distraction to what you describe (since you were able to find a path
with Zope). This is a demonstration that successful outcomes less
tightly bound to more traditional zope development can be achieved also.
:-) Many thanks.
Jeff Shell wrote:
So yes, it is possible to have good SQLAlchemy integration. But
'integration' may mean different things to different people. Some may
want invisible or near invisible integration with conventional Zope
content-management-esque concepts, integration with the ZMI, the
dublin core, etc. That's overkill for my needs, which is why I've
stayed away from those implementations. Having Location, Security,
Adapter binding (which yields views and URL traversal) is just about
what we need. The rest we provide through our own business logic as
Views, Utilities, and plain old Python classes and functions. And only
one new ZCML directive (yay!).
Zope3-users mailing list