Message-Id: Date: Fri, 22 Apr 94 11:49 BST From: Paul H J Kelly To: partain@dcs.gla.ac.uk Subject: Further details from J J Liu about our FEM code Liu has kindly provided confirmation that you're doing the right thing. I enclose his message. I am currently revising the paper he refers to. Paul ----------- >From nasoftwr.demon.co.uk!jliu Fri Apr 22 10:28:17 1994 Subject: Re: Finite element To: Paul H J Kelly X-Mailer: ELM [version 2.4 PL23] Content-Type: text Content-Length: 1551 > > re: "Cc: The Big Elephant " > > Please explain!! > > re: "Of all the versions in the tar file, is there one you would recommend > as "the standard thing written in standard Haskell"? (It's easier to > race HBC using identical source code.)" > > I have had a look around. I believe there are two candidates for a > benchmark version of what's described in the paper: > > Chalmers/fem/... > > and > > Glasgow/distr/... > ... ....... > > I do hope this helps. I have cc'd this to Liu in case he can help, but > I'm afraid it was quite a while ago, and he has other things to think > about nowadays. > If my memory serves me correctly, the only PROPERLY packaged HASKELL code is the one in Glasgow/distr/... . A paper corresponds to the code (except I/O part) is in ~jjl/document/report/fp-fem/paper.[ps|dvi] entitled "Functional Programming for Scientific Computation" submitted to ENGINEERING COMPUTATIONS (not included in the FAST collected papers). The paper did not mention I/O, since functional programming is not good at it at all. Though there are directories such as Yale and Chalmers etc, they were not fully tested due to the availability of various compilers at the time. All these directories exist, because I was trying to see how far various compilers could go with the FEM code orginally developed in Miranda and Hope+ (which was the origin of the report --- Functional Programming for Finite Element Analysis, included in FAST project collected papers). I hope this helps. Liu.