I know some of Thomas Papanikolaou's contributions (he was the originator of the LiDIA project) ended up (improved) in Pari. Some of Victor Schoup's contributions also ended up in NTL (or it may have been the other way around). And certainly there are your contributions which have ended up in mwrank (or vice versa).
Sadly however, there were 40+ contributors to LiDIA. Hopefully you are right and we can say get the number field code GPL'd. This might be a less difficult project. But the problem is that LiDIA is built in layers so that for example, the number field code relies on the Linear Algebra (LA) package, the Lattices (LT) package and the finite fields (FF) package. These rely on the simple classes, interfaces and kernel (which can be GMP) and the LiDIA memory manager. Thus it isn't possible to extricate the number field code as a *working* module. The best we could do would be to make use of the algorithms contained in LiDIA for which we don't need a GPL. We only need to know where these algorithms have been written down in published papers. There are I suppose the coding tricks which the LiDIA writers have used in their implementations of these algorithms, which might be useful to read. Bill. On 5 Sep, 14:04, "John Cremona" <[EMAIL PROTECTED]> wrote: > The LiDIA interpreter was an experiment really -- its development > stopped very quickly as the person who write it left. That was one > major problem with LiDIA -- most was written by people in Buchmann's > research group while they were there (as students or postdocs) and > that part was often not maintained well after they left. > > However, we don't need to have the whole of LiDIA under GPL surely: > we only want certain parts of the code to adapt for our purposes, and > so it might be possible to get the author of that part to release it > independently. (For example I wrote some small bits of LiDIA, but we > don;t need those anyway as they are also in mwrank). > > John > > On 9/5/07, Bill Hart <[EMAIL PROTECTED]> wrote: > > > > > > > A difficult, but not insurmountable, problem seems to be that each > > copyright holder would need to agree to GPL their code. The copyright > > holders do seem to be listed, and there don't seem to be too many of > > them. Some of them are certainly not going to complain, but someone > > obviously believed strongly enough that LiDIA should only be used for > > free for non-commercial purposes to add that condition to the license. > > So it is not clear to me whether LiDIA can be easily GPL'd, especially > > if those people still feel strongly about this. Extricating the > > contributions of individual contributors could be quite difficult > > given the enormous size of the package. > > > Something that I did not realise until recently is that LiDIA has an > > interpreter. > > > Bill. > > > On 5 Sep, 10:12, "John Cremona" <[EMAIL PROTECTED]> wrote: > > > I have been building LiDIA regularly over the years but not in the > > > last few months so (in particular) not with the latest gcc. I have > > > remained on the LiDIA mailing list though (which has been extremely > > > quiet for a long time) and know that people have been putting in the > > > patches necessary to make it build with current gcc. I'll look those > > > up and report back. > > > > About GPL-ing LiDIA, I will ask too since I was at one time listed as > > > a member of the LiDIA group, and know Buchmann reasonably well, > > > they/he might be cooperative! > > > > John > > > > On 9/4/07, Bill Hart <[EMAIL PROTECTED]> wrote: > > > > > I looked at the LiDIA list. Christoff continues to promise the new > > > > version is coming, this time before he goes on vacation. > > > > > Apparently the compile problems are to do with the new gcc. Some kind > > > > person has posted the changes they made to get it working at length on > > > > the LiDIA list. > > > > > I'll see if I can get it going and post timings for that too. I > > > > anticipate that all the different layers in LiDIA will make the class > > > > number code quite slow despite the algorithm being quite good. I will > > > > eventually code it from scratch myself, since this problem interests > > > > me greatly and fits in well with the quadratic sieve I've been > > > > writing. But obviously that's not going to happen overnight. > > > > > Bill. > > > > > On 3 Sep, 22:00, "William Stein" <[EMAIL PROTECTED]> wrote: > > > > > On 9/3/07, Bill Hart <[EMAIL PROTECTED]> wrote: > > > > > > > I tried to build LiDIA so I could do some timings. But it just > > > > > > doesn't > > > > > > build. It's a standard Opteron machine with G++ and GMP on it. I'm > > > > > > just doing a standard build, but it complains in ring_matrix.h that > > > > > > some function is missing. > > > > > > The last time I built LiDIA was maybe 1999, and it wasn't too hard, > > > > > though > > > > > it took a long time. Justin Walker used to also frequently build > > > > > LiDIA back > > > > > then. > > > > > > One issue with LiDIA is that I think -- and correct me if I'm wrong > > > > > -- that > > > > > it still officially has a non-GPL compatible license. There was a > > > > > mailing > > > > > list post from somebody several months ago that they > > > > > would GPL LiDIA, but I don't know if that person actually owned the > > > > > copyright, > > > > > or if the university would allow it, or if anything actually came of > > > > > it. > > > > > If not, it would be great is somebody who remembers could track this > > > > > down or email again to bug them. > > > > > > Having some benchmarks comparing PARI and Magma will be really useful. > > > > > But for now I'm mainly worried about even having some part of the > > > > > *functionality* > > > > > of Magma in algebraic number theory, since presently we are even way > > > > > way > > > > > behind there. > > > > > > -- > > > > > William Stein > > > > > Associate Professor of Mathematics > > > > > University of Washingtonhttp://www.williamstein.org > > > > -- > > > John Cremona > > -- > John Cremona --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~----------~----~----~----~------~----~------~--~---
