What I'd like to characterise as first generation hypertext theory is, first of all, that branch of hypertext theory that is 'critical' hypertext theory. In other words not the work of the programmers but of those who want and use hypertext systems and use their experience to provide or produce new theories of textuality, or reading, or cultural production (or whatever) on this basis.
These are people like George Landow, Michael Joyce, Stuart Moulthrop and Jay David Bolter. Their work, while exemplary and crucial, has originated from firm print traditions so that hypertext is contextualised, and understood, in its relation to the book.
Second generation hypertext theory is probably going to be the work of people like Aarseth where there appears to be considerably less investment (and anxiety) about "the book" and where digital "literacy" has always been present alongside analogue "literacy".
(My interest in video, cinema and hypertext is probably the other side of the coin of first generation theory.)