Gabriel Dos Reis wrote : > It probably is difficult because you have consistantly avoided to > tell me what the definition of equality of function is, so that I > check that the implementation does not something sensible. > I hope that (at least one of the) axioms could propose (a) a practical (b) determinist code, (c) not a decidable code.
(a) : very often this call will be over calculus (exact or float or expressions...) and this reduce will work for the extreme case for sum or product void. With my naive understanding 2 functions with the same name and the same signature are equal because the result is the same. We can suppose that axiom doesn't change the definition of a function between the left hand side and the right hand side of the equal in f=g (b) :I use too often maple and I really want a proved? determinist code. Either the = is always right or is always wrong, but never the test is fuzzy. It' almost impossible to find his bug when this feature is present... I don't understand how the interpreter is, but the problem with changing lisp vector must be surround. (c) I understand that the test can't be perfect (as tests over floats, equal for real, stream, series, and so.) So a partial equal like a eq in lisp will be enough. A perfect code should only recognize that the anonymous functions (x,y) +-> x+y, (x,y) +->y+x and +$TheRightDomain are equal. François ------------------------------------------------------------------------- This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone _______________________________________________ open-axiom-devel mailing list open-axiom-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/open-axiom-devel
