On Nov 24, 12:37 pm, William Stein <wst...@gmail.com> wrote:
> > (RJF) For example, claiming great advantages to rewriting working software > > in the language du jour (currently, Python). > > FUD. Sage does *vastly* more than "rewrite working software". I did not say that was the only thing that Sage does. And I have no problem with people writing new software to do new things using Python or any other programming language. And you know that. > > (WS) There are clear advantages to Python over Lisp that go far beyond > marketing hype. It is perhaps the world's most popular scripting > language, and certainly in the top 2 or 3. That does add genuine > value. Would you have promoted rewriting Lisp programs into Pascal, in 1980, when Pascal was popular? Regardless, there may be advantages to using Python for writing some programs that are more meaningful than popularity, if you are writing programs de novo. The argument tend to fade when someone hands you a fully working Lisp program, and you assert that there is a clear advantage to rewriting it in Python. Or having a summer project for a high school student to rewrite it. > > > It is further questionable if the claim is accompanied by accusations > > that the old program is slow, when it is > > technology like pexpect that is the bottleneck, etc etc. > > FUD. Maxima is slow at many things, because of very dated algorithms > and implementation issues. This has little to do with pexpect. I > wish Maxima weren't slow (say compared to Magma), since that would > make my life much easier. If the Lisp-language algorithm is dated, it can be updated. If the implementation of the language (Lisp) is at fault, then maybe it is because Sage has, until very recently, been using an inappropriate Lisp implementation (byte coded). There are certainly coding tricks that can make some algorithms fast that may be used in C, assembler, or Fortran, but are inconvenient in Lisp. That is, they may best be resolved by calling the program that was written in C, assembler, or Fortran. I imagine it is the same with Sage, Python, GMP, MPFR, etc. > >(WS) It's worth pointing out that Sage switched form using Maxima for > symbolic manipulation to (a massively modified version of) GINAC in > May 2009. I find it odd that you say you now no longer use Maxima for symbolic manipulation but you use Maxima for symbolic integration. What could you possibly mean? Do you mean a trivial simplification of x+x converted to 2*x is now done without Maxima? Woohoo! > http://www.ginac.de/ andhttp://pynac.sagemath.org/ > Maxima is used now only for certain specific capabilities not > available in GINAC, including symbolic integration, formal symbolic > summation, limits, etc. But it's not used for basic symbolic > manipulation anymore. Of course this is false, because symbolic integration uses a substantial amount of "basic symbol manipulation" and that is done in Maxima. So Maxima is used for basic symbol manipulation. Unless you have somehow replaced calls to the Maxima simplifer (a Lisp program) to calls to some GINAC program, which then would have to return Lisp data structures. > > > And it is > > also hard to give much credit to a system that advertises itself as a > > new viable alternative to Maple and Mathematica when the fact is that > > ^^^^^^^ > > The statement is "Mission: Creating a viable free open source > alternative to Magma, Maple, Mathematica and Matlab." Your assertion > is wrong on two counts: > > (1) There is nothing about "new" there. The word "creating" suggests that you are making something new. as .. "the act of making, inventing, or producing" http://www.merriam-webster.com/dictionary/creation Is this discussion productive? > (2) There is no claim that Sage *is* a viable alternative yet. > > > it often just calls Maxima, already a viable alternative to Maple and > > Mathematica, > > Maxima is not a viable alternative to Maple and Mathematica, for most > potential users. I really wish it were. Actually, for many users of Maxima, I think it is. There are some people (like me) who sometimes use Mathematica or Maple, but often I just use wxmaxima. You should try it sometime. > > > a reality that it attempts to conceal (see NSF proposal > > draft). > > I'm not hiding anything. > > > Until something goes wrong, and then it blames Maxima.(see > > Sage-support). > > FUD. We do not blame Maxima whenever anything goes wrong. No, you don't blame Maxima for every thing that goes wrong. Please pardon my hyperbole. > There are > over 800 bugs listed in the Sage trac database, and the *vast > majority* have nothing to do with Maxima. We "blame" Maxima when we > run into a Maxima bug, or a mistake in our understanding of Maxima. Right. > > It is worth noting also that actual Maxima developers (which RJF is > not these days, so far as I can tell) who have ever communicated with > anybody involved in the Sage project have been very _supportive_ of > the Sage project. I'm thinking in particular of Robert Dodier. It is possible to contribute programs and applications for Maxima without checking them in yourself to CVS. > > > I think that any company that chooses to use Sage instead of buying > > Mathematica would be unlikely to do so because Sage is open-source and > > can therefore be "verified". > > While Sage might be free, the expense of > > that company in hiring a person to check the verification would be > > substantial, and probably pointless. Proving that a program matches a > > specification or some other indication of correctness is not something > > done casually or cheaply. Verifying a program's result (whether from > > Mathematica, Maple, or Maxima) > > is typically done by quite other means than examining the code that > > produced it. I can recall no instance of a published program proof > > in symbolic computation for any non-trivial algorithm. Repeated > > proofs of the Extended Euclidean Algorithm (GCD) don't make the grade. > > The above is the typical kind of statement that comes form somebody > who doesn't understand mathematical research at all... combined with > the sort of statement that comes from somebody who doesn't understand > open source either. How many companies that buy computer software are in the "mathematical research" business? I can think of companies that do engineering research, financial research, physics research, biological, chemical, aeronautical, weapons, defense, .... . Mostly government, I think. How many mathematicians (that is, people whose title is "mathematician") are there in the US? According to the Bureau of Labor Statistics, in the most recent year for which they provide a number, in 2006 there were 3,000 mathematicians in the US. three thousand. see http://www.bls.gov/oco/ocos043.htm huh? how could that be? Well, BLS also mentions 54,000 "postsecondary mathematical science teachers". But anyway, it gives you pause to think how many people might be professionally or even academically concerned with, say, modular abelian varieties and the Swinnerton-Dyer conjecture, and related computer programs. > > * For mathematicians: Open source addresses the "proof" part of > research mathematicians curiosity. The connection between open source and proof of correctness -- or even the connection between written proof and correctness, is not as obvious as you might think. For example, many published proofs have been shown to be incorrect. What is the nature of a convincing proof? There is certainly a literature on program proofs, but even if a program were proved correct, how would you prove that it was executed flawlessly? (one of many questions that could be raised, even about open source programs.) Is there philosophically something, how shall we say, "sinister" about secret programs? Sure. But to say "open source" = "more reliable" is nonsense. > > * For companies: open source is something entirely different > altogether -- it's the power to know they can change the software > however they want. Do you think that represents much of a desire on the behalf of companies? > > For some people it's just nice to save money. Often open source > software is just solidly better quality and more functionality > software than anything closed source out there. Often? Really? It is my impression that there is a huge huge pile of really crappy open source programs, often attempting to do something that is nicely done by a proprietary program. There are a few open source quality projects, often driven by important economic factors or widespread use. Mathematicians are rarely in the driver's seat, in these respects. > It happens, and most > of us know examples that we personally use. Sure, what comes to mind is linux, mozilla, emacs, tex, maxima, gcl, gmp, mpfr, ntl. But I also use Maple, Mathematica, Microsoft Word, Microsoft .net, Microsoft speech SDK, tablet SDK, visual C++, Adobe Acrobat, Windows, Photoshop, Turbotax, Itunes, Skype, Allegro Common Lisp ..... > I can't say that I have ever puzzled over Photoshop, thinking that if I only had the source code I could rely on it more. 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
