Free math

[mmcirvin]

Many years ago, when I was a grad student, I bought a student-discount copy of Mathematica, the symbolic computer algebra system. It was crucial to finishing my doctoral thesis: I pretty quickly got beyond the point where I could safely do the algebra involved in my research by hand without making a fatal mistake somewhere. I knew some people who

Comments

[mmcirvin]

...Another option in Sage is to go a bit under the hood and send the function through its Maxima interface to Maxima's interactive viewer, which is a little pokey but looks OK.

[mmcirvin]

Fabrice Laussy on G+ pointed me to Mathics, a front end for SymPy and Sage that more closely imitates Mathematica. I haven't messed around with it much, but it's certainly pretty.

[oonh]

I use mpmath for most of my domain coloring work, though the maintainer has something new out called arb which I haven't seriously played around with yet. Also: things like mathematica and sage tend to be enormously too slow for the things I want to do with them (real time rendering of Bessel function Newton's method fractals and whatnot, and that's eventually going to require something very much like an asic or an fpga with special functions support).

I find mathematica really clunky in some ways: its not smart about breaking multi-line expressions in TeXForm, its numerics are abysmally slow in some cases, and it's a beast. On the other hand, some of the support for integration and specfun type stuff is probably more tightly organized than sage. (and that's just a matter of time, I think.

And yeah, graphics are *miserable*. I have some diagrams that I need to make and I'm looking at some ungodly mix of epix and tikZ to accomplish them.

[mmcirvin]

Mathematica's graphics capabilities were astounding and futuristic when it first appeared about 25 years ago (though, at the time, I was working on more advanced visualization at NCAR

[mmcirvin]

...and, yeah, Mathematica seems to beat everyone for special functions; that's probably its greatest relative strength.

[mmcirvin]

...I was just playing around with the Klein j-invariant in Sage, and it makes me appreciate how much crunching probably had to go into one frame of your little video about it.

[avitzur]

Conventions determined by user interface constraints set by line printer teletype terminals is depressing, I agree. At some point I'll retire and open source GC in an attempt to encourage the UI state-of-the-art into the 1980's.

[mmcirvin]

I think that with graphics, part of it is not so much line printers as laser printers: it's the assumption that what you're trying to produce is a printed figure for your paper, which means that scalable arbitrary-resolution graphics, bold black lines and easy reproducibility are more important than intuitive high-frame-rate interactivity or fancy shaders.

[oonh]

mathematica's scalable output is in some cases really poor: you get pdfs or svgs of 3d functions which are incredible unoptimized because they try to mimic the opengl process and they take forever to render in pdf viewers sometimes.(it would be nice if there was occlusion and clipping preprocessing).

[neowolf2]

Does the excellent free common lisp SBCL run on Macs? If so, it should be possible to get Maxima to run on it. Common Lisps have been pretty good recently about conforming to the CL standard.

[mmcirvin]

Maxima already runs on Macs: it's part of the Sage install, so if you have Sage you also have Maxima, and you can use it directly if you like.