[Haskell] Software is at Dead End?

Andrzej Jaworski himself at poczta.nom.pl
Tue Jan 30 18:22:10 EST 2007


Hundreds of cores will bring that level of complication that dumping things
onto declarative programming may turn out to be the only choice for the
software industry. I am surprised they haven't started to prepare for that
yet. Apparently their foresight remains the same as when their PhDs waited
for Bill to reinvent DOS;-)

Declarative programming however is not about glorious ideas but about
reducing them all together. For the average sequential Joe logic programming
is easier than lessons from aesthetics and thus simple tasks (e.g. without
arrays) will be easier for him to commit to OR-parallel Prolog. Then beyond
Prolog there is linear logic programming which has the same feel but brings
concurrency, constraints and agents by definition. Haskell should remain for
the chosen dozen:-)

Cheers,
--Andrzej



More information about the Haskell mailing list