I use haskell to to Project Euler problems, so my "project" directory looks like p1,p2, p3 ... pn for each different problem that I'm working on. inside each subdirectory there's a Main.hs and any other supporting .hs files.
I'm wondering if there's a way to use cabal with unix wildcards or regexes to build everything without needing an explicit executable section for each problem in the .cabal file.submitted by cptwolf
[link] [7 comments]
If I click one of the posted links to an external site (not the comments for that post) it appears on the sidebar under "Recently Viewed Links". How does Reddit know that I've viewed the external site?submitted by tomejaguar
[link] [7 comments]
Is it appropriate to ask for help writing an algorithm here in exchange for bounties (money)? If not, is there a better place?
So unless a mod posts that's it's fine, I won't place my specific inquiry here in order to not break the spirit of the post.
Basically I need a script that's beyond my ability to write and I'm willing to pay for it. It's not anywhere near a full on software project. I just need to process a set of values and run a (difficult) calculation on them and output the solution to the terminal or to a file or whatever. Based on my non-professional experience, I'd say it's on the level of an extremely hard homework assignment.
Is there a good place for me to ask for this? I'm asking in this sub, because the problem seems particularly suited to Haskell or other functional languages.submitted by rorriMnmaD
[link] [10 comments]
Anyone know a way to run a SIFT detector/descriptor from Haskell?
I've been doing my Machine Perception homework in Haskell so far, and don't want to stop here.submitted by CarbonFire
[link] [13 comments]
Wasn't one of Clean's original design goals to be an educational language for strongly-typed, lazy functional programming? So how come we're not hearing educators complaining that Clean's length function is overloaded and the one for  is just an instance?
In fact, I happened to go to a university where (at least at that time) Clean was used in the introductory functional programming class, and I specifically remember that when length was used in an early example, the lecturer mentioned that it's overloaded, we'll be covering typeclasses in a couple lectures, just pretend that its type is specialized to , and we swiftly moved on.
Why does the similar situation in Haskell suddenly freak out so many people?submitted by gergoerdi
[link] [44 comments]
I’m wondering if it would be possible to translate GHC Core or STG to Morte (calculus of constructions). More generally, is there any information/papers on the difficulties/limitations of Morte/CoC?submitted by altaic
[link] [26 comments]
Hi, all. Long time lurker, first time poster. Someone just posted about making a hex-grid game, so I'm finally coming out of the woodwork.
Could anyone look at the code and tell me about areas for improvement (or features they would like to see)? Feel free to attack anything... implementations, naming, module structure, testing, documentation, formatting, usability.
(edit: formatting issues, forgot link)submitted by josephwrock
[link] [7 comments]
I'm writing bachelor thesis about reactive programming in Haskell. I'm using Euterpea to develop simple audio application. I almost finished Haskell School of Music by Paul Hudak.
But, as some of you know, Mr. Hudak passed away earlier this year and the book is now finished - chapters on frequency modulation are only half way done. Physical Modelling chapter has a lot of empty spaces. I can't understand ideas presented there if there is no explanation :(
Is there any information whether the book is going to be finished? I don't even know how can I obtain such information.
Is anyone except Paul Hudak involved in writing Haskell School of Music?submitted by lovaz
[link] [6 comments]
We all love microbenchmarks, why wouldn't we?
So, there was a new microbenchmark posted, and again, it striked me that my naive Haskell implementation is twice as slow as python one ran from PyPy, and just as slow as Python one ran from CPython (regular "python").
Here's a link to my fork: https://github.com/k-bx/benchmarks/tree/master/json
- Generate json with ruby generate_json.rb (sorry)
- Run python-based one with time python test.py
- Build haskell one with cd json-hs && make && cd ..
- Run haskell one with time ./json-hs-bin
My results after 15 minutes of poking:➜ json git:(master) ✗ time pypy ./test.py 0.500162876712 0.499362626291 0.499902604092 pypy ./test.py 5.37s user 0.91s system 99% cpu 6.346 total ➜ json git:(master) ✗ time python ./test.py 0.500162876712 0.499362626291 0.499902604092 python ./test.py 11.44s user 0.81s system 98% cpu 12.443 total ➜ json git:(master) ✗ time ./json-hs-bin 0.5001628767123875 0.4993626262908803 0.49990260409169224 ./json-hs-bin 12.37s user 1.60s system 98% cpu 14.207 total
Original reddit post with discussions: https://www.reddit.com/r/programming/comments/3pojrz/the_fastest_json_parser_in_the_world/
So, my question is: is this "ok" or is there some relatively-simple trick that could be applied and considerably improve performance? Thank you.submitted by k-bx
[link] [25 comments]