RE: Getting started

Kent Landsfield sez:

> The ONLY reason that I've stayed with hypermail to date is that it is
> the
> fastest and impacts the system the least. It is far from perfect but
> interpretive converters require too much of the system to be able to
> effectively deal with large, high traffic lists. (This is being said
> WITHOUT ANY experience with pipermail).
>
> I have no problem with rapid prototyping with an interpreter (and do
> that
> often) but the end product needs to be as fast as possible and that
> points
> to a 'C' implementation.
>

Remember, too, that C is a lingua franca among computers. While I have a collection of programming languages on the various machines I work with, the only compiler that's on all my machines is C.

As soon as you get into other languages, even C++, you run the risk of it failing to compile or compiling erroneous machine code on one system or another, and of having to maintain object code for multiple platforms, OS versions, etc. (not that we might not *want* to have object distributions, but *having* to have them is another kettle of fish).

I have a favorite language or two myself, but if I'm distributing source code, I distribute vanilla C, and I make sure it compiles on gcc.

There's also the problem of redeveloping into the new language, be it Java, Python, Ada95 or whatever. If we want to get the accumulated patches into service as quickly as we can, I think we ought to work with what we have (C source) unless and until the accumulation of features makes a clean sheet of paper approach necessary. Then we can make the programming language choice. But I don't think we need to now.

Just an opinion from one of the folks in the gallery, of course.

Bob Crispen
bob.crispen_at_boeing.com Received on Thu 23 Apr 1998 06:21:42 PM GMT