Probably the best way to learn what an evaluation context is to read the book you've linked to there. Alternatively, it's basically what the Racket reference section on the evaluation model calls a "continuation": http://docs.racket-lang.org/reference/eval-model.html

We do try hard to make the docs understandable though, so if you have some feedback please do mention it on the mailing list so that the responsible devs can try to address it.