On Thu, Jul 09, 2009 at 02:47:37PM -0500, Jason Catena wrote:
>
> Yes, sorry I didn't look it up earlier.
>
> Bentley, J., Knuth, D., and McIlroy, D. 1986. Programming pearls: a
> literate program. Commun. ACM 29, 6 (Jun. 1986), 471-483. DOI=
> http://doi.acm.org/10.1145/5948.315654
[The article is reproduced in D. E. Knuth, "Literate Programming", CSLI
ISBN 0-9370-7380-6]
For the task to be done "print the k most common words in a file", the
Unix approach and the Unix tools give everything to create a "program"
far more rapidly than the from scratch approach adopted by D. Knuth. But
because the tools exist (are already written... but in what language?
Easily understandable? Maintainable? etc.).
But this does not mean that _in general_, literate programming has not
its strength especially for complex and weaven program... or even for
writing the tools, the bricks one combines in a pipeline like McIlroy does.
I don't think that TeX and METAFONT could have been written correctly,
or could be understandable in something else than WEB (unfortunately not
CWEB; that would simplify greatly "porting").
[For another thread: MetaPOST can be used instead of gnuplot. But not
easily for 3D like plotting. Unfortunately, MetaPOST too is WEB not
CWEB. And the ad hoc conversion of some Pascal to C (web to C) seems
alas the simplest way. Even the Pascal compilers that could be ported to
compile on Plan9 (if there are) would probably not allow a
straightforward compilation of the WEB based programs.]
--
Thierry Laronde (Alceste) <tlaronde +AT+ polynum +dot+ com>
http://www.kergis.com/
Key fingerprint = 0FF7 E906 FBAF FE95 FD89 250D 52B1 AE95 6006 F40C
