On Sep 23, 7:43 pm, kcrisman <kcris...@gmail.com> wrote: > On Sep 23, 11:19 am, rjf <fate...@gmail.com> wrote: > > > On Sep 23, 5:36 am, Burcin Erocal <bur...@erocal.org> wrote: > > > > I think it would be a huge overstatement to say that the symbolics > > > subsystem in Sage was "designed" in any way. IMHO, it was mostly > > > patched together to support educational use, then acquired more cruft > > > through several rewrite attempts and cramped schedules. > > > I think that is true of ALL the subsystems in Sage. > > As we have tried to convince you before, this is very much NOT the > case for large portions of Sage. In particular the sage-combinat and > much of the algebraic geometry/number theory piece are definitely > designed, certainly by this point.
I do not have any familiarity with those parts of Sage. If they were written specifically for Sage, they may play better. Or they may play better simply because the specifications are sufficiently clear that these parts do not require constant reconsideration. Just as Maxima would have very few surprises if you used it for (say) polynomial arithmetic over the integers. (Of course for that domain you could use any number of pre-built programs, or you could write your own in Python if you have bignums.) > > > In terms of my own interests, that is why the idea of rewriting Maxima > > in > > Python is pointless. > > But your interests are not quite the same as that of a practicing > mathematician who would rather *use* math software than create the > perfect one. Right. Yet the various proposed projects to rewrite symbolics in Python are not the same as those of a practicing mathematician, either. And neither is that of the practicing neuroscientist, > etc., who just wants easy access to something like our new Brian > experimental package along with some other mathematics. That's fine, > but not relevant in a normative sense. You have a neuroscientist using Sage? > > > Wouldn't you prefer to write a system that > > addressed > > the shortcomings of Macsyma (etc.) > > No. That would be a very good idea for someone whose research is > directly related to those shortcomings. Sage's Mission: Creating a viable free open source alternative to Magma, Maple, Mathematica and Matlab. > For most of those on this > list, we would prefer to contribute to a system that has > > 1) Killer app in web-based interface You think there is a killer app that involves (say) algebraic geometry? The closest anyone seems to have come to a killer app is education. (Probably pre-calculus). Possibly, "do my calculus homework", but that's more limited. > 2) Has cutting-edge combinatorics and algebraic functionality My impression is that people with large problems in combinatorics tend to write their own programs, and anyone else's program is either not computing the right thing, or is too general and hence too slow. This may easily be mistaken, since I'm judging from very few examples. > 3) Has access to huge amounts of other libraries I think that people do not generally want huge libraries. They want libraries with what they need. If they are forced to use a huge collection of probably irrelevant cruft, they balk. That's certainly my reaction to (say) "improved" Excel. > > That could have been Maxima, no doubt! But for whatever reason its > development went in other directions. That's fine. Maxima also wants > everything 'in-house'; that's fine too; That suggests you don't understand the state of Maxima, which imports large scientific libraries. BLAS, Lapack, Minpack, Slatec, quadpack. Slatec has about 24,000 lines of code. Sage prefers to use Python to > glue other high-quality projects together with its (substantial) new > code. Maxima tends to take a pretty-much working system and import code. This can be done by automatically converting fortran to lisp, or linking code in other languages e.g. as front-end or as plotting back-end. A better way, but not entirely portable is to use a foreign-function interface to load libraries. I use it frequently for accessing libraries written in C, C++, assembler... but only on the lisp I'm using, and not part of distributed Maxima. > > I still don't know why you don't consider the truly new code in those > areas to be interesting, but I suppose we all have the occupational > hazard of overspecialization; I certainly do. If there is new code in algebraic geometry, that's nice, but you are quite right that I don't find it interesting. And therefore putting it into Sage, or making it accessible only if you use Sage, or praising Sage because it includes such code, is pretty much irrelevant. > > > And by shortcomings I don't mean > > "written in Lisp instead of Python". > > Although this is a shortcoming in the sense that it is far more likely > that someone who is not a computer scientist by training might > actually be able to contribute, of course it isn't in the sense of how > to write a high-level symbolic program. I disagree. Anyone who can program can learn Lisp. Someone who finds Lisp too difficult to learn and instead decides to (re)write symbolic math basics in Python in order to contribute to the state of the art will likely (1) spend all the time writing basics and never advance the state of the art; (2) have an inferior program. > But that doesn't happen to be > a goal of this project, Seems to me that Stein was encouraging projects like rewriting Maxima in python, especially appropriate for summer-high-school students. > though of course we are grateful to the GPL > for allowing us to use the constantly improving symbolics in Maxima. > And if anyone with Lisp knowledge ever reviews Nils Bruin's patch to > make library access to Maxima the default, it will even be fast :) so > we won't complain about that any more. > > But please don't make these blanket statements about the whole content > of Sage until you are familiar with all those subsystems. Does that rule out using sarcasm? I suppose I will have to mark my comments, e.g. <sarcasm> I'm sure that Sage is the ultimate design for math </ sarcasm> > As I've > said before, and will likely say again :) Still, it's worth saying. > > All the best, > - kcrisman Have fun RJF -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
