[mtrycz2]

I also disagree with the premise that research happens on hypertexts.

When you do research you need to concentrate on the lecture at hand in-depth, not get lost in a web of references breadth-first style.

[jplayer01]

What? You need to survey hundreds of papers that are relevant to your research. Nobody has the time, nor is it expected or in any way efficient, to read even a majority of these in-depth.

[dredmorbius]

I read notes, bibliography, and index first. In that order. The more hypertext-like a work, the better, though it should be sufficiently self-contained that the references are supplementary.