Remove docstring for cdf(::Skellam, ::Real)

[devmotion]

IMO this should be an internal comment and not a public docstring. For users, there's nothing special about cdf(::Skellam, ::Real), so there's no need for a separate docstring. See #1985 (comment) for an example of the situation on the master branch.

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 86.28%. Comparing base (46bb0c3) to head (92e7886).

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #1986   +/-   ##
=======================================
  Coverage   86.28%   86.28%           
=======================================
  Files         146      146           
  Lines        8787     8787           
=======================================
  Hits         7582     7582           
  Misses       1205     1205           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[andreasnoack]

I think this makes sense although I have slightly mixed feelings. I really like that the R docstrings contain information about the source of the implementations. However, I think it will generally be harder in our case since most of the implementations aren't even in this package. They are in StatsFuns and they then again rely only functionality from SpecialFunctions and HypergeometricsFunctions.

[devmotion]

Another argument for this PR is that it seems surprising to single out Skellam - in case we want to provide more details regarding the numerical algorithms/approximations, we should do it consistently for all distributions. However, currently Skellam seems to be the only one with a dedicated cdf docstring.

Generally, I think it's a bit questionable whether docstrings are the best way to communicate these details though. Doing it for all distributions would mean that ?cdf is basically useless. Maybe a separate page in the docs would be a good way.

[mmikhasenko]

For the record, I support documenting technical details of the implementation.

It could be gradually implemented for all user-faced functions. It's useful and already easy to scroll to the right docstring in Pluto if there are many.The display can also improve one day to show only the Base.method and type specific method.

But it could be that Distribution.jl, by a consistent convention, decides not doing that.