From: "Andrew P. Lentvorski, Jr." <[EMAIL PROTECTED]> On Feb 3, 2005, at 1:30 AM, Gabriel Sechan wrote:
Doesn't matter what the problem space is.
Ahem. Did you even *read* my email. I gave a perfect counterexample.
Yes I did. I disagree.
The only reason I can see anyone claiming fortran is better than C, Java, insert other language, is that it has an amazing wealth of libraries for it that you may not be able to find in $language. That can influence your choice, although if you know those libraries chances are Fortran is your language of choice anyway. While I don't know too much about fortran, when you get down to it math is math. Write a function do the math and call it. C and C++ have very rich math libraries themselves, comparable to fortran.Numerics in Fortran.
Fortran has very explicit rules about pointer aliasing that makes it preferable to practically *every other language* with respect to Numerics. There is no way around this within the framework of other languages.
How about writing a device driver? I would be *insane* to try and write that in Python rather than in C. The API's pretty much lock me into that.
I think you're insane to write python anyway :) But assuming you have a Python compiler (not just an interpreter), there's no reason why you can't. The compiler just needs to make the ABI for the interface functions match the expected one of the OS.
Sure, there are lots of problem spaces where the language really does not matter. And lots of people tend to debate language rather than just get to the problem.
However, there are some places where it does. The trick is to know when you have one of those.
Which is never.
Gabe
--
KPLUG-List mailing list [email protected] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
