Thanks for your comments.

You are absolutely right to remind me of the GSL. The issue is that not everyone has the GSL, that it is overkill for many problems, and that it can be a bit of a challenge to install. A small, pure Perl implementation for the family of Gaussian functions only still seems desirable. (To me.)

In regards to the maturity of the algorithms: I am taking mine from Abramowitz/Stegun. That should be a sufficiently solid source.

I'll give more thoughts to your suggestions regarding documentation and naming. Much as I hate overly verbose names, you are probably right that a more descriptive name would be beneficial.