Re: Programming skills required by mathematicians?

On Jun 28, 8:59 am, Aatu Koskensilta <aatu.koskensi...@[EMAIL PROTECTED]
> wrote:
> Marshall <marshall.spi...@[EMAIL PROTECTED]
> writes:
> > The less educated are now screaming the "Java is slow" mantra that
> > dates from 1997, when it was true.
>
> The main objection to Java is of course not that it is slow but that
> it's awful.

Heh. I have to chuckle in recognition, but I also have to disagree a
bit.

There are languages such that, at some point when one is learning
them, one has a sudden revelation, and the room seems brighter,
and for a moment one feels smarter, and one may even laugh
out loud as sometimes happens during the reading of a particularly
elegant proof.

Java is not that language.

However, if someone entrusts me with the task of taking
a team of a dozen or two diverse programmers, ranging from
just-out-of-college to seasoned Ph.D.s and getting them to build
a server according to a set of halfway-written specs with strict
performance requirements that has to go live in six months,
I'm likely to pick Java.

As a language for the solo intelligentsium, Paul Graham style,
Java is perhaps awful. As an industrial language that can
absorb whatever new team members or capricious requirements
are thrown at it and scale up to a million users and a million
lines of code, it does quite well. This is no small achievement.


> > For the OP's question, namely for someone intending a math
> > education, it strikes me that there are only two good answers,
> > SML and Haskell, and I'm astonished that no one has mentioned
> > either one yet.
>
> Astoni****ng indeed. Of the ML family Objective Caml also deserves
> mention.

I did actually mention OCaml, albeit perhaps too briefly for what
it deserves. For pedagogical uses, I'd tend to say SML first;
for industrial ones I'd say OCaml, but perhaps I'm not entirely
justified in doing so.


Marshall