This is my first post to python-dev, so for those who might not know
me, I'm the author of Pro Django and more recently, Pro Python.
I haven't looked at the plugin landscape in a while, but I was very
disappointed the last time I looked at how complex typical systems
were in this regard. There seems to be entirely too much time spend
worrying about type checking, interface requirements and the like.
Python is based on simple ideas providing simple solutions, and I
would argue that duck typing is a perfectly acceptable route to take.
I'd suggest that any plugin system proposal start off simple and work
up as needs arise, rather than trying to cover as many bases as
possible at the outset. I know it doesn't cover all the bases yet
(particularly with regard to discoverability), so I'm not throwing it
in as a proper suggestion, but I'll point out that I wrote up a
dead-simple plugin system a while back[1] that may provide a useful
starting point for discussion.
I haven't read the existing proposals yet, but I just wanted to
breathe a word of caution into this before it gets too far into
interfaces and whatnot. I'll be glad to help with any discussion that
takes place on this, though. It'll be a good way to ease into the
mailing list, since it's an area where I've already spent a good deal
of thought.
-Marty
[1] http://martyalchin.com/2008/jan/10/simple-plugin-framework/
___
Python-Dev mailing list
Python-Dev@python.org
http://mail.python.org/mailman/listinfo/python-dev
Unsubscribe:
http://mail.python.org/mailman/options/python-dev/archive%40mail-archive.com
