[bitc-dev] BitC surface syntax

[Pierre THIERRY]

Scribit Kevin Hammond dies 08/08/2006 hora 11:37:
> [S-expressions] are certainly convenient for meta-programming, [...].
> But for ease of coding and good abstraction, a decent syntax is a
> great advantage.

I'm not sure the LISP history would back that statement. In fact, the
S-expressions along with macros precisely made possible to use a
``decent syntax'' instead, without modifying the language itself. There
were even a quite big number of such attempts, and the LISP community
finally rejected all of them (except for some very restricted
sub-languages, like the loop macro in Common Lisp).

Maybe the reason was totally irrational, like the desire to reamin some
sort of elite using a syntax unparsable by mere mortals, or some sort of
masochism. But the claimed and observed reason is that S-expressions
proved to their users to be the best syntax for ease of coding and good
abstraction.

Even Mc Carthy himself didn't see the S-expressions as a practical
syntax and wanted to use M-expressions for the real task of programming.
But very early it happened to be easier to program with S-expressions.

On this basis, I'm not sure it would really be a great decision to drop
S-expressions for the sole reason that they are an awkward and
unpractical syntax to work with, and because people complain about that.
I'm sure that when they really begin to do some coding in a sexpr BitC,
they will just love it.

There is another reason, though, as, as Shapiro pointed out to me
recently, the use of macros may be undesirable in BitC, and one of the
very reasons to ever use S-expressions is the strong adequation between
syntax and data structures holding it, these data structures being made
easy to manipulate (thus the natural early addition of macros to LISP).

So if BitC code won't be stored and manipulated as nested singly-linked
lists, maybe S-expressions are just useless[1].

Carefully,
Nowhere man
 
[1] If we don't take in account the fact they are probably the best
syntax ever ;-)
-- 
nowhere.man at levallois.eu.org
OpenPGP 0xD9D50D8A
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://www.coyotos.org/pipermail/bitc-dev/attachments/20060814/00f97548/attachment.bin