William D Clinger scripsit:
> To what extent does the push to add new procedures
> to the R6RS, including procedures that are simple
> compositions of other R6RS procedures, derive from
> fear that these procedures would be too slow if
> Scheme programmers were to write them themselves
> or to use a portable reference implementation?
Speed's only one such factor: convenience is another, correctness
(it's hard to roll your own and get all the details right, like with
hash tables) a third, and the correctness of larger programs
(the motive for Common Lisp's ENDP) a fourth.
> In other words, how much of the growing size of
> the R6RS can be blamed on the Scheme community's
> habit of using interpreters instead of compilers?
Is that really true?
--
If you understand, John Cowan
things are just as they are; http://www.ccil.org/~cowan
if you do not understand, cowan at ccil.org
things are just as they are.
Received on Mon Jun 25 2007 - 19:09:42 UTC