erf(0) should return 0

[sagetrac-rossk]

Currently it doesnt...

sage: erf(0)
erf(0)

---

Apply:

• attachment: trac_8983_v3.patch to $SAGE_ROOT/devel/sage

CC: @burcin @kcrisman

Component: symbolics

Keywords: erf

Author: Benjamin Jones

Reviewer: Burcin Erocal, Douglas McNeil

Merged: sage-5.0.beta5

Issue created by migration from https://trac.sagemath.org/ticket/8983

[burcin]

comment:2

Here is the relevant thread on sage-devel:

http://groups.google.com/group/sage-devel/t/d236e80bf7826bff

The main problem is this:

sage: erf(0)
erf(0)

We should just return 0.

[burcin]

Changed author from RossK to none

[burcin]

Changed keywords from erf zero to erf

[sagetrac-dsm]

comment:5

Beginner reporting for duty! So I have a patch which sets erf(0) to 0 -- or, rather, to parent(x)(0) if x==0, so that it's the "right" zero; I think this is the right idiom -- and adds complex arguments from mpmath.

But I have a few questions that came up along the way:

(1) In 4.6.1, solve(erf(x)==0,x) seems to work for me already, before the patch. Did I get my binaries confused somehow, or did something change elsewhere?

(2) It'd be nice if #8720 (fixing str(CC(0))) were finished, it surprised me when writing tests for the zero type preservation. Maybe I should have a look, I think it's mostly fixing some doctests left.

(3) My patch broke a doctest in functions/special.py, which expected erf(0.5) to be evaluated to 0.520499877813047, but the existing doctests for erf demand that erf(2) = erf(2). I can get that behaviour, namely to evaluate for reals and to hold for integers except for zero, but I don't quite follow it.

(4) Since I'm doing a bit more, should I open an enhancement ticket instead?