On 2014-12-05, Fredrik Johansson <fredrik.johans...@gmail.com> wrote: > ------=_Part_1821_492426932.1417781508772 > Content-Type: multipart/alternative; > boundary="----=_Part_1822_1964954233.1417781508772" > > ------=_Part_1822_1964954233.1417781508772 > Content-Type: text/plain; charset=UTF-8 > > On Friday, December 5, 2014 8:17:44 AM UTC+1, Nathann Cohen wrote: >> >> Helloooooo everybody ! >> >> I am preparing some Sage talk, and I wanted to say at some point: >> "Honestly we are not that good. We have strong points but we miss many >> things too. It all depends on what the developpers are interested in: we >> are great on some research areas, and under water level on others" >> >> Somehow this question is also related to William's "Sage has failed", as >> we cannot be a replacement for Mathematica/Maple/.... unless we cover all >> kinds of mathematics. >> >> In your past experiences (possibly when using Sage to teach in a >> classroom), in which areas do you think we are behind users' expectations ? >> > > * Symbolics in general. Symbolic integration and summation. Rewriting > expressions. Simplifying inequalities. Simplifying expressions subject to > assumptions involving inequalities. Limits and generalized series > expansions involving special functions. Sage can do a bit of these things > via Maxima and SymPy, but it's nowhere near as powerful as Mathematica. > > * Many numerical functions do not work with arbitrary precision. In > Mathematica and Maple, arbitrary precision works seamlessly pretty much > everywhere. In Sage, a lot of functions are hardcoded for double precision. > A first step would be to provide really solid support for basic numerical > calculus (like computing integrals, #8321) by leveraging what's available > in Pari and mpmath. When it comes to things like multivariate optimization, > solving ODEs and PDEs, etc. with arbitrary precision, I don't think there's > any open source software that competes with Mathematica.
The latter is not that uniformly bad. E.g. Sage beats the MMas on arbitrary precision linear optimisation :-) Asn well, there are OSS's, not (yet) in Sage, that do arbitrary precision semidefinite optimisation (a particular kind of convex optimisation, generalising linear) something that the MMas don't do. Dima > > * Performance. You often run into a wall as soon as you do anything > slightly complicated in Sage that isn't directly wrapping the right > C/C++/Cython implementation. Bill Hart had an example of computing a > resultant of two large (but not astronomically large) multivariate > polynomials over a finite field, where Magma does it in a minute, Bill's > Julia code does it in 5 seconds, and he had to kill Sage after waiting for > hours... > > Fredrik > -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
