[plt-scheme] Re: xxx chooses MzScheme as preferred language

> They were right though I have heard of Scheme implementations that are
> supposedly competitive with C++ but nobody has ever seen them at work
> ;-) Some compiler writers promise a lot.
>
>> However, I must say that other than this kind of development with
>> severe efficiency needs, PLT Scheme is the way to go.

As a new schemer, here are the three issues
that I see are significant in the programming world
when choosing a language (ignoring library support) -

1. Abstraction ability
2. Design rigour
3. Efficiency

(not in any particular order)

I find lisp/scheme (any lambda calculus) good for (1) with a
C-interface supporting (3). A rich type-system like the Hindley-Milner
one in Haskell helps with (2), though in no way complete for the task
(need range checks, invariant checks, etc.) MzScheme's contracts help
with (2) as well.

Lisp is awesome for me when it comes to playing with DSLs 'cos
I can write macros and play god. Its nearly perfect as a general
data format as well. Just interface with C for that cool
signal processing routine that you need.

If its not happened yet, I expect multi-lingual system development
will become the norm some day. Why constrain oneself with the question
"can this language help me in all aspects?" Lets ask
"which languages help the various design problems
that I'm facing?" instead. There is often more than one these days.

Regards,
-Kumar