On Wed, May 20, 2009 at 1:24 PM, Jose Mario Quintana
<[email protected]> wrote:
> I do not recall making the statement that someone in particular was 
> "suggesting J is weak or in-efficient in
> solving 'functional programming' problems."

I was the one who indicated that somebody said that, and the statement
I had in mind was made elsewhere. Brian Hurt, an Ocaml programmer,
posted the following comment to Twitter on May 9th:

] If a language doesn't support tail call optimization, it can't claim to
] be "good for functional programming". This isn't necessarily bad.

I thought (and continue to think) that this claim is inaccurate. I
wanted to challenge the claim in public, and I wanted to improve my
understanding of why it is inaccurate. (The discussion that has
subsequently occurred here has been both beneficial and gratifying to
me.)

> ...
> The first paragraph was mean to convey that from my function-level 
> programming vantage
> point (which seems to be close to Raul Miller's and away from Tracy Harms' 
> and your [Bill
> Lam's] 'functional programming' viewpoint):
>

My preference and intent is to work from a function-level vantage
point. True, I've been bringing into this conversation things that I
have been finding in the wider "functional programming" community, but
those are often things that I don't understand or like so well as what
I've come to use in J. This particular project began with the intent
of reading the advice of programmers who write in "other functional
languages" then apply those ideas in J. What I've learned is that
although there are very important lines of continuities, there are
also some stunning divergences of concern. (Raul's recent posts
explain this phenomenon to some degree.)

How can we participate in function-oriented programming discussions,
which are multiplying rapidly and gaining prominence, while avoiding
the confusions that arise if the depth of difference between J and the
alternatives is not appreciated? I'm not expecting an answer to emerge
quickly, but I do want to place the question before us.

As I put myself up against examples and discussions of the
better-known "FP" languages, I'm slowly getting a better sense of how
J compares to and contrasts with them. Soon I hope to write some
comments about how it looks like function-level programming has less
to do with tacit programming than I used to think, and more to do with
the consistency that comes with bivalent syntax, as opposed to the
open-endedness of the lambda calculus.

> ... Sure, J is not a typical functional programming language; then again, J 
> is not a typical programming language.
>
> I just hope and guess we all agree that, regardless of the personal 
> programming style, J is for joy!
>

Absolutely!

Tracy
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to