Closed progval closed 9 years ago
Sympy Gamma fail in the same way : http://gamma.sympy.org/input/?i=diff%28cos%29
It is perfectly coherent. Indeed, if we want derive ln, the only correct answer is x \mapsto \frac{1}{x}. So, we need a new variable and a functional notation, so a system which contains lambda-calculus. So we have a problem for higher order function or variadic which are overloaded. For instance, D(log) must be derivated as the function x \mapsto log(x) (decimal logarithm) or as the function (b,x) \mapsto log_b(x). Any initiative would be crappy.
The only thng to do, is to return nothing.
I don't understand how this is different to input diff(cos(x))
; you still have the domain issue.
Error reported to Sympy Gamma. https://github.com/sympy/sympy_gamma/issues/58
Because we have a variable 'x'. diff(cos) is a function -> boum; diff(cos(x)) is a variable -> no problem. We do not have to define a functional notation. We do not have ambiguous cases (diff(log(x) cannot be understood as the function (b,x) -> log_b(x)). There are many differences!
Marc CHEVALIER
Le 18/01/2015 21:08, Valentin Lorentz a écrit :
I don't understand how this is different to input |diff(cos(x))|; you still have the domain issue.
Error reported to Sympy Gamma. sympy/sympy_gamma#58 https://github.com/sympy/sympy_gamma/issues/58
— Reply to this email directly or view it on GitHub https://github.com/ProjetPP/PPP-CAS/issues/10#issuecomment-70423455.
tan
is a function too, and it is not defined on an point of R. So would be diff(cos)
.
It is not the problem.
Marc CHEVALIER
Le 18/01/2015 21:14, Valentin Lorentz a écrit :
|tan| is a function too, and it is not defined on an point of R. So would be diff(cos).
— Reply to this email directly or view it on GitHub https://github.com/ProjetPP/PPP-CAS/issues/10#issuecomment-70423682.
On input
D(cos)
, I get this traceback: