Jim Fulton wrote:
Raphael Ritz wrote:
Thanks for this clairification, Jim. Alan, does that address your concerns? Any reasons left, not to adopt the five approach to Zope 3?
Just understand that the Five approach is still being developed, so there's nothing to "adopt" yet. :) But I certainly encourage folks to participate and help Martijn figure out what the approach should be.
There's the 'approach' and the implementation. The approach is fairly clear: a focus on baby steps to integrate into Zope 2.7. The aim is to introduce as much as possible as make sense of Zope 3 facilities into Zope 2. The implementation is still flux. Though that said, things like interfaces and adapters should be stable enough (as long as Zope 3 is stable in that regard), as there's really no difference between the way Five does them and the way Zope 3 does them. For views, we're moving along nicely following the Zope 3 pattern. At the post-Europython sprint Stuart Bishop and I, with help from Jim, came up with a way to integrate them the right way for Zope 2 objects. I'm now pretty happy with them, though some more possibilities and details are bound to change still. You can now even add views to existing Zope 2 objects, without the /edit/ hack, from ZCML. That this happens by way of "Structured Monkeypatching" shouldn't concern anybody. ;) While Five is in flux, I expect the main changes you'll have to make in your applications is slightly altering some ZCML statements and possibly changing an import here and there. It also depends on who contributes what, as always. :) Regards, Martijn