#haskell - Sat 19 May 2007 between 00:17 and 01:37



dons?src concatMap
lambdabotconcatMap f = foldr ((++) . f) []
skewhi dons
sioraiocht?src Maybe (>>=)
lambdabot(Just x) >>= k = k x
Nothing >>= _ = Nothing
sioraiocht?src IO (>>=)
lambdabotm >>= k = bindIO m k
sioraiocht@type bindIO
lambdabotNot in scope: `bindIO'
sorearJohnMeacham: Are you aware of the GHC Trac?
dons?users
lambdabotMaximum users seen in #haskell: 340, currently: 282 (82.9%), active: 25 (8.9%)
donsoh, 340 is a new high score. nice.
soreardons: When will the fusion team be done with nobench?
donsthe results got wiped by accident. i'll rerun them during the week, i suppose.
after a jhc rebuild.
( i accidentally redirected the fusion results to the wrong file)
i'd imagine in the longer term i'll just update the results every few months.
or on demand,
sorearHow much space do the nobench results take?
I'm wondering if it would make sense to do a cron.weekly pull build and run, with indefinite historical results
donsspace, not much.
its just a matter of writing infrastructure
WeremanateeHaskell books put a lot of emphasis on proving programs. Do haskell programmers do that in practice to any extent?
ddariusYes for various values of "prove". Though most don't, I would say. However, most use libraries that have that approach applied to them.
donsWeremanatee: semi-formal methods are usually used, related to proof techniques. some automated provers for particular properties are also used.
BinkleyI'd imagine the most common proof technique in practice is "it typechecks, so it works" :-)
donsfor some commerical and research scenarios, proofs are indeed produced, (e.g. L4 /Haskell kernel, or various galois.com products)
ddariusI was about to say, you can get a lot of "proof power" out of the type system alon.
donsand QC is a semi-formal method too.
ddariusAnd (ab)using the type system in this manner is pretty common.
QC was one of the values of "prove" I was thinking of.
WeremanateeWhat is QC?
donsWeremanatee: so the general idea of reasoning about programs in terms of proofs is pervasive.
ddarius@where quickcheck
lambdabothttp://www.cs.chalmers.se/~rjmh/QuickCheck/
KorollaryQC is about disproving of course, but it's got inductive stuff.
ddariusOther program transformation results get used every day silently by the compiler, e.g. fusion.
donsalso though, the process of stating QC properties is similar to stating properties to prove.

Page: 3 10 17 24 31 38 45 52 59 66 73 80 

IrcArchive