PEP 266: Optimizing Global Variable/Attribute Access

[Alex Martelli]

"Peter Caven" <p.caven at ieee.org> wrote in message
news:_lGe7.43578$O8.6622552 at news1.busy1.on.home.com...
> Thanks Alex,

You're welcome!


> I sometimes forget how completely dynamic Python actually is!

...or "was" (although it will be for a while longer) -- performance
issues may well dictate a gradual (and hopefully only partial) loss
of some of that little-used dynamism, as forcefully argued by
Guido (and a few supporters) in a couple of recent threads.  At
some point there will be classes you cannot change, and that will
give higher performance when you explicitly tell the compiler
those classes are not dynamic (or maybe, it's still possible, the
non-dynamic will be the default, and you'll explicitly have to ask
for backwards-compatible dynamism instead).

Interesting times...!-)


Alex