What do you think of Mathematica?
How strong a general purpose tool do you find it for scientific work?
What do you think of the language?
Why I ask: I have been wondering lately about getting back into it. 10 years ago I used it to model some aerosol science and filtration theory from textbooks. I remember symbolic and numeric math, the “Notebook” interface that blends graphs and documents with code that gets evaluated, and (at least I thought) more consistency throughout the code than I experienced with Waterloo Maple 15 years ago. Since then I have done much more work with the SAS System (SAS Institute, Cary, NC). Some of my work is statistical, and I used canned regressions and the like. But an awful lot of my work is generally numerical, maybe with graphs.
I find more and more often I am using SAS to call other programs, feed them text and take other text back from them. For example lately it has been computational fluid dynamics with Fluent and Gambit, and finite element analysis with FlexPDE, but this year I’ve also written interfaces to CAD and .DXF files, to .SPC files with Raman spectroscopic data, to Garmin’s proprietary GPS files, and for reading text streams from lab instruments like precision thermometers with RS232 ports. Is SAS a good choice for these things? It works, but I just kind of fell into it.
I hear, for example, that some people have done a little CFD work directly in Mathematica.
Is it a reasonable base system for doing things like calling other packages and encoding and decoding binary and ASCII data?