|
Re: The Charlwood Fifty / Sage and sympy
Posted:
Jun 14, 2013 10:44 AM
|
|
Something like sympy can only be described at best at a particular time/stage of development. My guess is that the spectrum of what the integration part does can be described as roughly this.
1. a stub of a program with name "integrate" 2. easy-to-do integrals for demo purposes mostly 3. more serious rational function methods added 4. partial implementation of Risch "algorithm" 5. retrenchment of easy-to-do integrals upon realization that simplification, differential algebra, and other algorithms matter 6. partial implementation of Meijer G-function transformations. 7. Re-implementation of more of Risch "algorithm" 8. (for Mathematica) marketing claims of doing everything in book X, even if false. 9. adding pattern matching with a vengeance ala Ruby. 10. Reorganization, go back to step (1).
Where in the spectrum the current program sits, I don't know. It is of course possible that sympy will have the best integration program there is. But considering the amount of code it is alleged to require in competing implementations, one wonders how much effort will be needed to match it. Of course if one of the Sage methods is "call Fricas and Maxima and sympy in parallel and see what comes back" Sage might be interesting, if kludgy.
|
|