[Python-3000] Fwd: Conventions for annotation consumers (was: Re: Draft pre-PEP: function annotations)

[Paul Prescod]

Okay, you're the boss. The conversation did go pretty far afield but the
main thing I wanted was just that if a user wanted to have annotations from
framework 1 and framework 2 they could reliably express that as

def foo(a: [Anno1, Anno2]):

All that that requires is a statement in the spec saying: "If you're
processing annotations and you see an annotation you don't understand, skip
it. And if you see a list, look inside it rather than processing it in some
proprietary fashion."

It kind of seemed obvious to me, but I guess everyone's ideas seem obvious
to them. There were other secondary things I would have liked but this
seemed like the minimum required to protect programmers from "greedy
frameworks" that don't play nice in the face of unfamiliar annotations.

On 8/16/06, Guido van Rossum <guido at python.org> wrote:
>
> There's much in this thread that I haven't followed, for lack of time.
>
> But it seems clear to me that you've wandered off the path now that
> you're discussing what should go into the annotations and how to make
> it so that multiple frameworks can coexist.
>
> I don't see how any of that can be analyzed up front -- you have to
> build an implementation and try to use it and *then* perhaps you can
> think about the problems that occur.
>
> Collin wrote a great PEP that doesn't commit to any kind of semantics
> for annotations. (I still have to read it more closely, but from
> skimming, it looks fine.) Let's focus some efforts on implementing
> that first, and see how we can use it, before we consider the use case
> of a framework for frameworks.
>
> --
> --Guido van Rossum (home page: http://www.python.org/~guido/)
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/python-3000/attachments/20060816/b8481a6c/attachment.html