Re: linux headers and C++

Marc Mutz (Marc@Mutz.com)
Tue, 06 Jul 1999 17:10:11 +0200

Frank Butter wrote:
>
<snipped double quoted for brevity>
>
> until now I thought (and experienced) exactly the opposite.
> C++ is imho especially perfect to maintain bigger projects and
> was designed for this. to make programmers life easier in bigger
> projects and development teams. in a well designed C++ project you
> are actually forced to do the right thing (e.g. use inherited
> objects in the right way - and yes, use the overload feature
> to simply add features to an inherited object works great
> and does not confuse at all, if you are using it in the
> right way...).
> so do you really think C++ (and OOP in general) is a
> retrogression compared to C ??
>
Frank,

you can't possibly mean to write greater parts of the kernel in C++?
That language is just at too high a level to do that poking in hw
registers and designing time critical paths with. You need assembler and
C for that. Compilation results must be foreseeable as most as they
possibly can. If you look at the tweaking that is done in time critical
areas of the kernel (look for 'odd looking' goto's), what would that
exten(d/t?) to if c++ were used? You can even try using Java for the
kernel. You have too little control over the code generation, and if you
can gain it somehow, then the tweaking necessary will be _much_ uglier
than it is now.
I believe low-level software should be written in a low-level language.

Marc

-- 
Marc Mutz <Marc@Mutz.com>                    http://marc.mutz.com/
University of Bielefeld, Dep. of Mathematics / Dep. of Physics


PGP-keyID's:   0xd46ce9ab (RSA), 0x7ae55b9e (DSS), 0x31748570 (DH)






-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/