Stackless 3.0 alpha 1 at blinding speed

Tim Peters tim_one at email.msn.com
Sun Apr 20 18:51:41 CEST 2003


[A.M. Kuchling, points to
 http://mail.python.org/pipermail/python-dev/1999-July/000467.html
]

[Anton Vredegoor]
> Thanks. It seems Tim is making a heroic effort to explain the new in
> terms of the old. This necessarily must fail, but what other option is
> there instead of rebuilding all terminology out of gotos?

You're missing a decade of history <wink>:  as that post says at the start,

    Here's my biased short course.

I argued in favor of adding generators to Python starting in about 1991.
Every time I tried, the effort got hijacked by people going on about
continuations.  The post's primary point was that generators *can* be
explained "in terms of the old", and easily.  I think it succeeded at that.
Coroutines are harder, but can still be seen as a natural extension.
Continuations require a different mindset entirely, and I believe that post
was instrumental in (finally) getting generators taken seriously as a
feature that didn't require buying into continuations first.

Continuations are a different *foundation* for implementing control-flow
features, more than they're a feature in their own right.  While they're an
elegant foundation, they're not the foundation Guido (or 99.9+% of the
world's other language designers) chose at the start, and it remains
unlikely that any Python Guido works on will adopt them.






More information about the Python-list mailing list