#haskell - Fri 9 Mar 2007 between 05:06 and 00:12

NY Lost Funds



bitwiseshiftleftah. yeah, that's kind of big
malcolmwdons: just wanted to check, are you OK with sorting out the GSoC application sometime before Monday?
ulfdozhm, fw. checking my inventory.
monochromYou still think less and code less. You just code: "a[i] depends on a[i-1] and a[i+1]". Compare with an eager language: you have to hardcode the proper order of evaluation too.
zeeeeemonochrom: in my case, you think more and code more. after staring at the program and being unable to figure out where the violation is coming from, you keep changing and writing new code to try to slice up your haskell program to debug it.
editedheaven!
zeeeeefurthermore! i wrote this in the comments, but i believe an imperative solution is actually more natural for this particular algorithm, because then we would avoid these 'undefined' cells (Nothing).
dibblego:)
edited, you'll notice no silly stuff in here :)
monochromYou can code an imperative solution as you see fit.
editedindeed
zbrownzeeeee: that may be true, but in many cases a functional solution may be more elegant ;)
sjanssennewsham: how long did your solution to the "Programming performance" problem take to execute?
zeeeeei thought for a while why this was the case. we have data dependencies that are selective, and we're trying to express that with 'pulling.' however, it would be more clear to express the propagation/flow of data in this case using a 'push' model.
actionzbrown wishes that linux would quit dumping core files every few hours
zbrownwishes that linux would quit dumping core files every few hours
editedeven linux needs to take dumps
its healthy
zeeeeethis then made me realize that FP is 'pull', and imperative programming is 'push'
monochromI prove my programs correct, and I know haskell well enough to reliably do so. I am denied access to certain hard-to-hunt coding errors.
zbrownedited: true... i just need to write a cron job that hourly goes through dirs and delets core.*
editedgood idea
zeeeeethis realization agrees very much with my other perspectives of FP's relative weaknesses (e.g. can't write a program that counts the number of occurrences of each word - why? because this would be most easily/efficiently expressed with *pushes* to hash tables)
faxwhy create these core files if you only delete them?
dibblegofax, we should have lazily evaluated core files?
zbrownfax: i don't create them, system does them
faxhaha
zbrown: well its your computer
zbrownfax: eh, silly linux ;)
dibblego"hi I am your operating system, I have a core file behind this thunk here if you'd like to invoke it"
faxo99999999p0p-[p0-=][p0o;'09po['iolk;,/.op[
kfishor you could just disable core files if you don't want them: http://aplawrence.com/Linux/limit_core_files.html
lambdabotTitle: Controlling core files (Linux)
actionzbrown suspects lazy evaluation of core files would result in no core files being dumped if the evaluation basis was on core's usefulness
zbrownsuspects lazy evaluation of core files would result in no core files being dumped if the evaluation basis was on core's usefulness
kfish: interesting, hadn't thought to just remove them from being made ;)
kfishit's linux, you're in charge dude
faxyeah and also the idea of digging a hole and then filling it in 10 mins later is jsut mad
lispya trick i've used in hp-ux where it creates core files whether you want them or not, is to touch core; chmod a-w core. Then nothing can create a core in that directory becaus it exists and isn't writable :)
zbrownlispy: haha nice
monochromI am using Linux and I don't get periodic spontaneous core dumps. I submit that your setup is wrong.
zbrownmonochrom: what distro you using?

Page: 4 11 18 25 32 39 46 53 60 67 74 81 88 95 

IrcArchive

NY Lost Funds