I was going to do a proper blog post to interrupt my silence; unfortunately I’m having difficulty in getting the TeX to work properly.  In particular, for some reason the formulas don’t appear in the right place.

Anyway, this is material that I learned from a course of Sinan Güntürk; once they’re ready, I’ll post the notes I take. 

What I’m posting here now states that signals whose frequencies are bounded can be recovered entirely from sampling at a discrete set of points provided the sampling mesh is sufficiently small.  This fact was popularized by Shannon in his fundamental paper on information theory.

Here is the file.