On 26/08/2008, at 12:42 PM, William Stein wrote:
> 1) Let me clarify -- I *wrapped* the pattern matching already in > Pynac. ... > and you > can read about Ginac at the Ginac website: http://www.ginac.de/ Thanks for the clarification. Sorry I missed the post about pynac--- I'm fairly quick to delete stuff. In the spirit of clarification, I should say that I am not particularly batting for Lisp, Maxima or Mathematica, nor do I want to make sage look like Mathematica. But this area of pattern matching is one where the Mathematica / Lisp way of doing things is extremely instructive. That is the only reason I am rabbitting on about it. >> sage: sin(1+sin(x)).subs(sin(w0)==cos(w0)) >> cos(cos(x) + 1) >> >> In Mathematica, the result would be cos(sin(x) + 1), since having >> matched >> the outer expression, the replacement algorithm moves on rather >> than in. >> One would obtain cos(cos(x) + 1) by making the substitution rule >> sin==cos. >> Can you do that? Can you evaluate code during the matching of your >> patterns? Can you nest patterns? I think these are highly >> desirable, even >> if they are suitably carefully hidden from the casual user. > > I didn't implement the pattern matching. I wrapped the pattern > matching > that is provided by GiNaC. It is certainly not identical to the > pattern > matching in Mathematica, at present. I'm glad you're pointing out > some > of the differences between what is in Mathematica and what is in > Ginac. > Thanks! I see. (Sorry for sounding like I was jumping down your throat.) I was actually asking what capabilities you envisage. From looking at ginac, I think there's a good chance that it does all those things. >> We already have a pretty good "symbolic engine" in Sage, maxima does >> quite a good job of solving integrals, limits, etc. The main problem >> with Maxima is that we cannot extend it. >> >> Is "not extending of Maxima" a concrete policy? > > Maxima is definitely not the future of symbolic manipulation *in > Sage*: I don't know maxima well enough to have opinions on that, (except that it looked like a good idea at the time.) > (2) Maxima is written in Lisp. This is a total show stopper for > most > Sage developers, including me. In the spirit of "learning a new programming paradigm makes you a better programmer" I strongly recommend this page: http://www.defmacro.org/ramblings/lisp.html In any case, ginac might just be a good, non-buggy, reimplementation of just that part of lisp we need, and one that is more efficient for the problem at hand. >> I understand that maxima >> sucks in some circumstances, but it seems quite the beast here. >> I am quite confused about a lot of the pattern matching >> discussion. AFAICT, >> that is the problem for which lisp rocks, and the best way to do it >> is >> probably to write maxima code for it. And the best way to make it >> accessible from python is to provide pretty python methods to call >> lisp. >> Any other suitably general form of pattern mathematical matching will >> involve the proverbial reimplementation of lisp. > > That would be a waste of time for Sage. No matter how good Maxima > is or its pattern matching is, Maxima is still working in it's own > world of objects > and data types, which have nothing much to do with Python's. Maxima/ > Axiom > live on a totally different planet than Python/C/C++. This is just > a fact > we have to deal with. I actually don't understand this. For both pynac and the existing maxima link, I think you must already have some method of transforming a python object to and from lisp-like expressions. I suppose it might be problematically inefficient. At any rate, I am not after implementation details at the moment, sorry for bringing that furphy up. D ================================== David J Philp Postdoctoral Fellow National Centre for Epidemiology and Population Health Building 62, cnr Mills Rd & Eggleston Rd The Australian National University Canberra ACT 0200 Australia T: +61 2 6125 8260 F: +61 2 6125 0740 M: 0423 535 397 W: http://nceph.anu.edu.au/ CRICOS Provider #00120C --~--~---------~--~----~------------~-------~--~----~ 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://www.sagemath.org -~----------~----~----~----~------~----~------~--~---
