On 04/27/2013 08:36 AM, Jerzy Karczmarczuk wrote:
Christopher Howard: Is the portability which worries you, or the age of your system?
Actually getting a successful build and installation would be great. Also, there are multiple systems I work with, both of which have ancient software, but unfortunately are not the same configuration. I often find software that builds on one, but not the other.
Hugs (and Gofer before) are simply sufficiently old... I used them on Red Hat in one of my previous lives. Do you really need to compile your system from sources?
I guess not, if I can get one to install successfully to a local (non-root) user account. As mentioned, GHC Linux binaries failed me here, because apparently the gnu libc version is too old. With most software, I generally have had more success installing from source than trying to work with pre-built.
There are binaries everywhere. If you want a *simpler* language, perhaps try Miranda? Also a quite ancient language...
Or, perhaps a newer one, in some aspects simpler than Haskell (but far from any simplicity): Clean.
To be clearer, I do not really want any language other than Haskell. I just imagined that a simpler language might have a simpler and more portable compiler.
Perhaps it might help to know what do you need it for...
In brief, I have access to some large super computer systems. Sadly, nobody in my academic or work circles seems to have the slightest interest in applying functional languages to parallel computing problems (C and Fortran seem to be the languages of choice.) So, I've been poking around with some functional languages, trying to see what I could get installed (without any admin assistance whatsoever) and how I might be able to use them with the MPI or even GPGPU infrastructure we have. But I keep running into problems, because the software infrastructure is quite ancient (for compatibility purposes, I'm told), or there are other mysterious configuration issues. -- frigidcode.com