Meta: PEP discussion (was Re: PEP 255: Simple Generators)

[Tim Peters]

[Steven D. Majewski]
> I think one of the bug missing pieces to make it fit smoothly into
> Python was iterators.

I sure hope the "bug" part there was a Freudian slip <wink>.  Iterators are
indeed the bee's knees:  the existence of the wonderfully minimal two-method
iterator protocol answered most of the endlessly vexing questions about how
generators should act.  "Well, for starters, they should conform to the
iterator protocol.  Hmm.  There's really not much more to it than that, is
there?!".  So separating these concerns was one of those beautiful unplanned
accidents, falling out of a mere decade of intensely focused preparation
<wink>.

> ...
> Back in those ancient pre-PEP days, Guido was much more conservative.

I'm not sure about that!  I'd say more cautious in some ways, out of concern
that a small misstep could kill off his creation.  But after a decade of
never being wrong (and now paying me to tell him that <wink>), I think he
has more *confidence* in predicting the consequences of language design
decisions, so is less hesitant when he's sure he's right.  But Python is
still introducing new keywords at the rate of one per 5 years ...

> And there were fewer people to bounce ideas off of --- I recall I
> was soliciting ideas for a more pythonic syntax, but there wasn't
> much interest. ( And the above hack was even too ugly for ME to
> defend! )

Me too <wink>.