usage/future status of PyProtocols

[Tim Peters]

[rjt-usenet at thegrindstone.me.uk]
> We are considering making use of the PyProtocols
> (http://peak.telecommunity.com/PyProtocols.html) package. I have read
> through the documentation and it appears to offer much of what we
> want.
>
> I have already used both Zope's

Zope or Zope3?  They differ too.

> and Twisted's interface implementations to a greater or less extent in
> existing projects and I have been a bit disappointed that no 'standard'
> approach to interfaces has appeared in the Python libraries.

There are two PEPs open in the general area, 245 and 246.  You can read
Guido's most recent public thoughts about them here:

    http://www.python.org/doc/essays/pepparade.html

You may infer from the tone that Guido isn't overly excited about either of
them -- but of course that would just be your opinion <wink>.

> Before jumping into PyProtocols I was wondering:
>
> Are other people using it in real projects? What are their experiences
> regarding payback for effort expended?

I don't know, but I suggest also/instead asking on the PEAK mailing list:

    http://www.eby-sarna.com/mailman/listinfo/PEAK/

> Does anyone know if there are plans to include a 'standard' approach
> to interfaces/protocols in the Python library?

That's what the PEPs are about.  They haven't been rejected yet, so
"someday" I expect they'll make progress.

> If so, is it likely that it will be PyProtocols?

Your guess is as good as anyone's.  I'm pretty sure Guido doesn't use any
interface or adapter gimmicks today, so he won't favor one based just on
familiarity (there was some chance that Zope's mechanisms would gain an
advantage on that ground when Guido was employed by Zope Corp, but no more).

> I just don't want to spend a lot of time, and persuade my team
> members to use a package that is then superseded by a different
> approach in a future standard library package.

Sorry, no guarantees here.