PEP 245: Python interfaces

[Tim Peters]

[Martin von Loewis, on
 http://python.sourceforge.net/peps/pep-0245.html
]
> Even though this is quite a long document, it is incomplete.

[Michel Pelletier]
> Yes, *sigh*.  I know I need more formalism, but should this be at the
> expense of the section on goals?  I'm not sure.

Shoot for "in addition to" rather than "at the expense of".  The formal
specification sections needn't contain any motivation at all (indeed, are
probably better if they don't!); contrarily, the motivation sections can
defer to the specification sections for details.

> I was raised in an XP environment, the goals are the most important
> part to me.  This is something I've been debating myself over.  I
> obviously want the PEP to be accessable and not boring, but I also
> want it to be comprehensive.

To the extent possible, a PEP should be complete and precise enough to allow
someone from Mars to implement it, and leave you happy with the result in all
respects.

Don't forget that we're implementing a programming language here:  what the
PEP specifies has to cover all possible cases, and even the tiniest
advertised details have to remain stable for years and years to come.  Like
it or not, it's very rare that we get to do any form of refactoring in a
language or library spec once released:  people rely on the details within
hours of getting it.  Heck, look at the person-decades of turmoil caused by
the mere *suggestion* that Python make a trivial change like having 1/2
return 0.5 instead of 0 <wink>.

*these*-goals-can't-be-met-until-the-details-are-precise-ly y'rs  - tim