[Interface-dev] [Fwd: Re: Interfaces status and coordination]

[Glyph Lefkowitz]

On Thu, 2004-06-03 at 15:24, Glyph Lefkowitz wrote:

> I want to format methods for calling from the web, provide schemas for
> database persistence, verify constraints on remote method calls, and
> provide serialization instructions for remote method calls.

After talking with Itamar a bit about these use-cases, I think that the
right thing to do is to subclass InterfaceClass and do my
experimentation with different methods of specifying there.  Are there
any design problems with this approach?

I do want to see, and hopefully eventually standardize upon, the spec
you come up with, Jim, but I'd like to move Twisted (and Quotient, and
Nevow) away from our old, cruddy  component system implementation ASAP,
and I don't want to introduce extra dependencies.  (Getting the Twisted
slackers into shape for a release is a big enough challenge as it is.
^_^)