#apache #archlinux #asterisk #centos #debian #gentoo #haskell #kde #kubuntu #lisp #math #mysql #perl #python #ruby-lang #rubyonrails #suse #ubuntu #vim #wikipedia 0 1 2 3 4 5 6 7 8 9
Top Prev 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 Next
#haskell
<allbery_b> you might try someplace like #mysql
<christine_c> ok thanks
<iron32> Thanks a lot for all you help I got it working
<nmessenger> \o/
<iron32> and Improved my Knowledge of monads too
<iron32> :)
<nmessenger> meh. A monad is just something that can be (f)mapped over, joined together, and returned into. It's really not so much scarier than that.
<iron32> well not everyone is a Haskell master ;)
<siti> :D
<mbishop> only dons
<mbishop> and cale
<mbishop> ;P
<nmessenger> I've just found that when I go to *use* a thing with a scary name, I find there's usually not really much there to be afraid of.
<iron32> Modelling IO with monads might be the most logical thing but it sure isn't the easiest
<sorear> ... back
<allbery_b> enh. it's not really that bad
<allbery_b> most of the time you can lift stuff out and work purely
<nmessenger> the following page *really* helped my understanding of the IO monad in particular, I recommend it:
<nmessenger> @wiki IO_inside
<lambdabot> http://www.haskell.org/haskellwiki/IO_inside
<iron32> thanks I will look at it
<allbery_b> heh
<nmessenger> a function takes some parameters and results in a value, so if it depends on or changes the outside world, logically, the outside world should be one of its parameters.
<nmessenger> the only sensible thing you can do with the real world is destructively update it, which you can do purely functionally, so long as you never re-use an old world.
<Excedrin> it's interesting that nhc98 is always faster than yhc, and that hugs is faster than either in one case; I guess jhc fails to compile most of the tests?
<nmessenger> the IO monad as explained on that page is just a world-parameterized state transistion function.
<nmessenger> (though, as some have mentioned, its not the necessarily the *best* way to implement the IO type, but it can give a semblance of understanding)
<Excedrin> I guess "20% faster than nhc" on the yhc page is based on some other set of benchmarks
<allbery_b> really, you could do it by having every IO function take a final argument which is some non-instantiable (without hackery) value
<dons> it was at one point, but they've introduced some slow downs.
<sorear> Excedrin: much older benchmarks. you should read the ML, ndm was as suprised as you are
<dons> Excedrin: yeah, jhc fails to compile or run any of the tests :-)
<SamB> dons: typical
<sorear> dons: where does it say jhc failed?
<Excedrin> dons: based on the darcs jhc?
<sorear> dons: results.html doesn't list jhc
<Excedrin> sorear: it was mentioned earlier on irc
<thorat> dons: $ ./build; spits out "scripts/BotPP.hs:41:25: Not in scope: `B.breakChar'"
<thorat> ... when building lambdabot on debian etch ghc 6.6
<dons> thorat: grab darcs version of lambdabot
<thorat> k
<iron32> Anyone here use HARE ?
<thorat> I will
<thorat> vim and haskell, a match made in heaven
<iron32> thorat : It only supports Vim ?
<thorat> uh, dunno
<thorat> but I will know ;)
<sorear> dons: looking at your latest patch 'more stuff', it appears it will fromJust: Nothing if passed "jhc". I take it this means jhc isn't being included?
<iron32> So far I am using Eclipse as my Haskell IDE but I would like to have the ability to refacotr code
<dons> sorear, well, that's a quick hack.
Previous Page Next Page