[Python-3000] Draft pre-PEP: function annotations

[Paul Prescod]

On 8/12/06, Phillip J. Eby <pje at telecommunity.com> wrote:
>
>
> I mostly like this; the main issue I see is that as long as we're
> recommending best practices, we should recommend using tell-don't-ask (via
> duck typing protocols, adaptation, or overloaded functions) so that their
> libraries can be enhanced and extended by other developers.


Would you mind suggesting text for the PEP as an addendum to what I
proposed? And an example of both bad and good practice?

>"In Python 3000, semantics will be attached to the following types:
> >basestring and its subtypes are to be used for documentation (though they
> >are not necessarily the exclusive source of documentation about the
> type).
> >List and its subtypes are to be used for attaching multiple independent
> >annotations."
>
> I'm not sure why we would use strings for documentation, but I'm not
> opposed since it eliminates the question of multiple interpretations for
> strings.


I don't understand your point. Is there a better use for strings? Or a
better type to associate with documentation? Or you just don't see a need
for inline parameter documentation? The PEP itself used string docstrings as
an example.

>(does chaining make sense in this context?)
>
> I don't know if I know what you mean by "chaining".  Good use of
> tell-don't-ask means that any interpretation of annotations nested in
> other
> annotations would be defined by the enclosing annotation (or in an
> overload
> for it).


Yes, it's clear what nesting means. I'm not asking about nesting.

The question was whether there should be any relationship implied by the
fact that an annotation appears to the left or right of another annotation
in a list of annotations.

def a(b: [doc('x'), type('y')]): pass

Is there any sense in which the function 'x' should be passed context
information that would help it wrap or communicate with 'y'?

The most likely answer is "no" but function decorators do chain so I just
wanted to raise the issue in case anyone wanted to make the case that
parameter and return code annotations should as well.

 Paul Prescod
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.python.org/pipermail/python-3000/attachments/20060812/31cb1cab/attachment.html