iterators (was: python-dev summary)

Steve Holden sholden at
Wed Feb 28 15:38:30 EST 2001

"A.M. Kuchling" <amk at> wrote in message
news:slrn98rkib.ik.amk at
> On 16 Feb 2001 18:55:27 GMT, Donn Cave <donn at> wrote:
> >If it's a big step towards something useful, fine.  I have heard
> Agreed.  Some time ago I argued that once Unicode was added to Python
> 1.6, we should just stamp the language definition "done" and leave it
> alone.  Sure, we can add little tweaky features like iterators, but
> will they enable a startling new application domain?  Will they
> attract more users?  In both cases the answer is likely "no", so we
> should just write off the new features -- they might be nice, but why
> bother making books obsolete and making the code more bloated?
> I think I'll repeat the argument for 2.1: call the language as defined
> by the 2.1 docs the final version, and leave it alone.  Obviously you
> could still work on a faster interpreter, fix bugs in it, etc.  This
> means interfaces won't get implemented, which is unfortunate, but if
> you let in "just *one* more feature" it'll never stop.
No, really, have just one more feature -- they're wafer thin!

Apart from leacing a lot of pretty competent programmers with less to do, I
can't think of many disadvantages to freezing the specification. This, I
presume, was originally the thinking behind Py3K.

It seems to me that the future may cause the Python applications base to
slowly crumble, as incompatible features which "break very little code" are

but-then-i'm-just-an-old-fart-luddite-ly y'rs  - steve

More information about the Python-list mailing list