Yes this is what I was looking for and somewhat expecting. I can understand the challenge in getting ZClass, Py/Perl Script, "composite components" done right. In fact to me they could easily make things more complex. Potentially more complex process and more complex machinery to handle it. I would definitely be for a simpler model of components which are just as describe .py files and an interface. I read somewhere that doing components right is/can be hard. The author wrote that component creation is done by the few for the many. There will be fewer component authors and more component users and integrators. I agree with his assessment. I think what will make it easier for component authors will be well a defined API. This appears to me what y'all are doing. If I understand correctly. Thanks. Jimmie Houchin Michel Pelletier wrote:
On Tue, 27 Feb 2001, Jimmie Houchin wrote:
While reading the Zope Development Roadmap about components I had a question.
It says: """Components will be edited via the filesystem as .py files. Components will probably be checked into and out of Zope via a CVS like facility. Components can be tested locally without checking them into Zope."""
What does this say about developing components with Py/Perl Scripts? It looks to be closer to the current Python Products.
Yes. We have thought a bit about "composite components" and "persistent modules" and stuff like that, but we only went so far into elaboration when we realized that it required lots, and lots of thought and effort.
The current component effort is much simpler, a component is: an object with an interface. If this is a ZClass with Perl-based script methods, then so be it, but we haven't thought far enough into what ZClasses really are to start thinking about giving them interfaces.
Is this sort of what you're asking?
-Michel