Can anyone recomend a good intoduction to C...

[Alex Martelli]

"Marcin 'Qrczak' Kowalczyk" <qrczak at knm.org.pl> wrote in message
news:slrn9aiaq7.rmu.qrczak at qrnik.zagroda...
    [snip]
> > > systems for other languages, time-critical modules for other
languages.
> > > C++ doesn't fit there, is unnecessarily too complicated for these
> > > tasks, and it isn't a high level language either.
> >
> > C++ fits *perfectly well* for these tasks.
>
> Maybe under Windows, but it's rarely used at all under Linux.

The Boost Python Library works perfectly well under Linux,
and so does Mozilla and the attendant XPCOM component
model.  They're C++ babies.  Writing such stuff, that fits
perfectly well with your model of tasks for which C++ is (you
claim) "unnecessarily too complicated", is made a LOT more
productive by using C++ rather than C.

> Kernel is in C, system interface is in C, system-level libraries are
> almost exclusively in C. C++ is used in some projects (KDE, nsgmls)
> but lower level libraries are generally expected to be available in C.

"Expected" is correct; it's a social issue, NOT a technical one --
technically, C++ is a better choice for most Python extensions &c
on a Linux platform.  But, if one wants acceptance, one writes in
C (e.g., that's what I did for GMPY, because I'd like people to use
it -- but it would be already out of alpha if had dared use C++, and
it would be just as fast &c).

Your opinion is widespread, and self-perpetuating, but it has no
current technical basis for such tasks as Python extensions on Linux.


Alex