>> Many thanks. Is there a documented version available? > > Martin, when I looked at that code, I was asking myself the same question. I > guess Mike might answer: the documentation is at > http://www.risc.uni-linz.ac.at/people/hemmecke/aldor/combinat/index.html > I find that very unsatisfying. I've put *a lot* of energy in the > documentation which probably caused that Mike was able to implement that > stuff so quickly. Now, if in the future we would like to transfer some ideas > from sage-combinat to aldor-combinat. We are in a very bad situation. That's > somehow unfair and I hope that people working on sage-combinat will do a bit > more than writing doc-strings. Mike, just imagine I had not documented the > way I deal with 'approximateOrder'. I guess you would have had very hard > times to understand my code.
Yes, there were places where the documentation was very helpful. There were also places where the actual Aldor code was the most helpful. > Where is all your background information and your design decisions. You > probably could not follow the Aldor approach exactly. Where do I find a > documentation of the different design? I guess my first priority was to get code out for people to play around with and to get feedback on the design since it is definitely not set in stone. Code is much easier to refactor than documentation. I'll write more big picture documentation when I know that things have settled down. Why types of questions would be the most useful for you to have answers to? > So, before I say yes, what does that actually mean? And why do you need it? > You are not using Aldor-code anyway. > > My position is actually as follows. AC is under GPL2, because GPL3 was not > available when we started. Now I am not really against adding "and later > versions" to our GPL license. But first explain, why you need that. If you don't view my code as a derived work, then there is no issue and nothing needs to be done. If you do view it as a derived work, then I can't release it under GPL v2 or later (which is the license for Sage code since it is the GPL license compatible with the most software). You could release a copy just to me under v2 or later, and I wouldn't distribute your code under that license. You could leave the version on the web under GPLv2. > >>> I'll probably be working on the multisort case later on. Due to lack of >>> funds, I can't make it out to RISC this summer, but I'd be more than >>> happy to >>> join in your discussions about multisort species. >> >> I did most of the maths involved (i.e., how to generate isotypes of a >> composition of multisort species), it's mainly a problem of >> representation. >> The trouble is, so far I did not find a reasonable way to do composition >> such >> that the output are *representatives* of isomorphismtypes -- the approach >> taken >> in the trunk version of Ralf's code and mine. But if your code above >> solves >> this problem, then there is little left to do! > > Mike, actually, I'd like to follow your code using mercurial, but I've no > experience with it. If you could give some hint on what to do exactly to be > always up-to-date with your latest sage-combinat and what to do to actually > run it. If you have a copy of Sage right now, you can download this script http://wiki.sagemath.org/combinat/Mercurial?action=AttachFile&do=get&target=sage-combinat, make it executable, make sure sage is in your current path, and run "./sage-combinat install". This will clone a copy of your sage repository, make one for combinat, and apply our patches there. The next time you start up sage, it will be with the combinat/ branch and have our patches applied. If you do "./sage-combinat update", it will update to our latest patches. To make them active, you run "sage -br combinat" There is a fairly thorough overview of our development process at http://wiki.sagemath.org/combinat/Mercurial > BTW, I haven't looked into Multisort species for some time. Theoretically > that is not too difficult. Why I don't go on with it is, because Aldor > doesn't let me do the definition as nicely as I want it. > > Basically, I'd like to write > > F(G,H,K) > > similar to what is done in the univariate case (see > http://www.risc.uni-linz.ac.at/people/hemmecke/AldorCombinat/combinatsu26.html#x40-580008.13 > > Actually that is not a problem for bi- tri-, ... n-variate species. > But that would mean n implementations of basically equivalent code. > As you might guess, I don't want to double code and rather write generic > code for that situation. Aldor doesn't let me specify this at compile time. > Since Python is interpreted, you might be more lucky. I see. So the issue is that Aldor classes/categories can't take variable number of parameters? I guess I avoid the problem by having my species be objects/elements instead of classes/domains/categories. --Mike ------------------------------------------------------------------------- Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW! Studies have shown that voting for your favorite open source project, along with a healthy diet, reduces your potential for chronic lameness and boredom. Vote Now at http://www.sourceforge.net/community/cca08 _______________________________________________ Aldor-combinat-devel mailing list Aldor-combinat-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/aldor-combinat-devel
