[Axiom-developer] Re: Limits in Axiom

[Ondrej Certik]

> Aha. Do you have lots of testcases like that one in SymPy?

I tested all the limits in Demidovich, you can find SymPy testcases here:

http://sympy.googlecode.com/svn/trunk/tests/test_demidovich.py

(it should be easy to understand it, if not, just ask me)

And some more testcases (from Gruntz) are here:

http://sympy.googlecode.com/svn/trunk/tests/test_limits.py

> I think one week is just too over-optimistic. Any new implementation in
> Axiom now requires to write in a literate programming style. So one
> actually has to understand the thesis first and then put that together
> with the program into a pamphlet (which is an article like form that
> contains the program). Maybe we should ask Gruntz whether we are allowed
> to include part of his thesis in an Axiom pamphlet then.

We chose a different approach for a development of SymPy:

http://www.extremeprogramming.org/

basically, if the tests run, the code si fine. Whenever we find the
code ugly, we just refactor it - we can do whatever we want as long as
all the tests run.

So we have tests for every functionality that we want SymPy to have.

Ondrej