Python is just as good as C++ for real apps

Alex Martelli aleax at
Fri Jan 25 11:08:03 EST 2002

"Courageous" <jkraska at> wrote in message
news:u3035ukscffh2i9sfklqpj9fvnf1a8h7vu at
> >Potential for extremely high performance has always been very high
> >among C++'s design objectives, even to the detriment of simplicity,
> >generality, and other nice concerns.
> Yes. Where C can be construed to be a "portable assembly language,"
> C++ can be thought of as a "portable assembly language with object
> oriented stuff thrown in and generics strapped-on as a later after
> thought." C++ might have been a different language than it is today
> if Bjarne hadn't had his hands tied by management.

Maybe, but my impression, from Stroustrup's "Design and Evolution"
book, is that these crucial design decisions were ones he made, not
ones Bell Labs Research mgmt (or any other mgmt) foisted on him.

> Which reminds me of a small curiosity. I read once that during the
> design stage, Bjarne apparently considered making it so that the
> statement:
> int* p1,p2,p3;
> ...would evaluate to...
> int *p1,*p2,*p3;
> I've always thought that this was proper. The compiler treats (int*)
> as a type, and so it should behave like one, in my opinion.

Maybe, but it would conflict with the target of "C with Classes"
being C-ish enough to allow very fast and easy porting of existing
C code (and each later step was constrained to compatibility with
the previous one as substantial production code existed).


More information about the Python-list mailing list