On Tue, Dec 16, 2008 at 05:31:11PM -0500, Gwern Branwen wrote:
> My intuition has always been, if it reduces parentheses, it's probably
> good. That is, 'foo a = baz (bar a)' would be reduced down to 'foo =
> baz . bar', which has fewer parentheses, and so is clearer. But
> something like 'f x y = z (x y)' turns 'f = (z .)' with the same
> number of parentheses, but is worse. (And some expressions can lead to
> horrifying pointless versions, needless to say.)
note that in the presence of 'seq', eta reduction/expansion is _not_
always a valid transformation. Furthermore, it can affect optimization
signifigantly if the compiler doesn't have enough info to eta expand
everything out.
John
--
John Meacham - ⑆repetae.net⑆john⑈
_______________________________________________
darcs-users mailing list
[email protected]
http://lists.osuosl.org/mailman/listinfo/darcs-users
