[Python-Dev] Python 3 optimizations continued...

[Nick Coghlan]

On Thu, Sep 1, 2011 at 3:28 AM, Guido van Rossum <guido at python.org> wrote:
> On Tue, Aug 30, 2011 at 10:04 PM, Cesare Di Mauro
> Cesare, I'm really sorry that you became so disillusioned that you
> abandoned wordcode. I agree that we were too optimistic about Unladen
> Swallow. Also that the existence of PyPy and its PR machine (:-)
> should not stop us from improving CPython.

Yep, and I'll try to do a better job of discouraging creeping
complexity (without adequate payoffs) without the harmful side effect
of discouraging experimentation with CPython performance improvements
in general.

It's massive "rewrite the world" changes, that don't adequately
account for all the ways CPython gets used or the fact that core devs
need to be able to effectively *review* the changes, that are unlikely
to ever get anywhere. More localised changes, or those that are
relatively easy to explain have a much better chance.

So I'll switch my tone to just trying to make sure that portability
and maintainability concerns are given due weight :)

Cheers,
Nick.

P.S. I suspect a big part of my attitude stems from the fact that
we're still trying to untangle some of the consequences of committing
the PEP 3118 new buffer API implementation with inadequate review (it
turns out the implementation didn't reflect the PEP and the PEP had
deficiencies of its own), and I was one of the ones advocating in
favour of that patch. Once bitten, twice shy, etc.

-- 
Nick Coghlan   |   ncoghlan at gmail.com   |   Brisbane, Australia