Malcolm.Wallace at cs.york.ac.uk
Wed Feb 8 13:16:08 EST 2006
Marcin 'Qrczak' Kowalczyk <qrczak at knm.org.pl> writes:
> In my language Kogut there are only pairs, and larger tuples are
> expressed by nested pairs (biased in the same direction as lists,
> without an end marker of course).
> I wonder whether the performance difference is really that
> significant. Short tuples seem to be much more common. Anyone could
> gather statistics about runtime usage of tuples of varying sizes?
I agree that most uses of tuples are small, and for these, a nested
representation will not be so bad. But Robert was proposing a
nested representation precisely for the case where the tuples are
large (possibly machine generated), and it is in those cases that
performance might start to matter.
More information about the Haskell-prime