Is there a way of linking the NAG library to an Algol68 program on a Ubuntu (Linux) system

Question

I'm re-visiting Algol68 because I think it's historically important. I would like to link one of my programs (written as a student) to the NAG library routine f07abf (linear equation solver). The question is, being a retired academic, do I need to purchase the NAG library, and how do I link it to the program using a68g on my Ubuntu system? The NAG website itself is quite technical and I need some guidance.

Answer 1

I don't have a full answer, but a few hints:

1. a68g does have available "The algorithm that GSL uses in LU decomposition is Gaussian elimination with partial pivoting" cf. learning-algol-68-genie.pdf and GSL LinAlg. This GSL version might make a useful alternative.
2. Also: a68g can generate (and compile) C-code (on the fly) and then load+call the resulting C library.so ... It would depend on whether you have the band-width to figure out this C-code generation/compile/load/call process and reroute via FORTRAN NAG library.
3. I have previously contacted someone at NAG, they hinted that the classic Algol68 version of NAG has been lost, they also hinted (if someone can find a copy) that they might be reasonable about giving permission to use the Algol68 version of NAG. I suggest you find your old university's version, and then ask NAG for permission to use. (Let me know what version you do find, your university may have one in archive, ask them, it might be on disk archive, or still on a dusty old BASF magnetic tape, tapes can still be read.)
4. Pick a different implementation of A68... eg. You can call "ALIEN" subroutines from algol68toc (aka Algol68RS), so you can call a FORTRAN or C version of f07abf. I've done somethings similar and it worked.
5. BTW: I believe Paul McJones @ Computer History Museum (in Mountain View, California, US) would be happy to catalogue any Algol68 NAG version you find. That way others can follow in your footsteps.
6. There also is a sometimes-active group for algol68 at linkedin where you can ask these kind of questions.
7. Also: I believe that a lot of the early NAG routines were borrowed from non-NAG authors. So you might like to track down the original f07abf author for permission to use the original f07abf.
8. re: "re-visiting Algol68 because I think it's historically important..." That's why groups like HOG existed in 2003, and will still continue in 2103... ;-)

I hope this helps.