review divergence, footprint, and resolution

Update webdocs.

• Improve "beam divergence" example: use sample with highly structured GISAS pattern to show the effect of the divergence [Redmine2349].

API:

• Remove remnants of old-style divergence definition from API [Redmine2421, reported by DY].
• Provide divergence for OffSpecular simulation in Python API [Redmine2300, reported by GP].

Is there current need for resolution function from data or from free model function? [Redmine2422]

Support Monte-Carlo sampling over parameter distributions [GitHub913].

Also covers Remine2484 ("unify footprint correction") and Redmine1102 ("review normalization, especially in view of finite sample extension, finite beam width, and small glancing angles")