In his Hypertext ’97 keynote address, broadcast from the US to Southampton, UK, Ted Nelson lamented that (for the hypertext community) the Web was like waking up one day and discovering you had a teenage son, who was a juvenile delinquent. There was much concern among hypertext researchers that the web in general (and Tim Berners-Lee in particular) had gotten it all wrong because they had ignored the years of scholarship about how to build hypertext systems.
This sentiment is rooted in the view of hypertext as structure. The ad hoc nature of the web, the lack of link integrity, the lack of separation of link structure from content, all were used to critique the way the HTTP and HTML were used in conjunction with the web browser to implement hypertext.
If one looks and hypertext as behavior, on the other hand, the web presents a different picture. The premise that you can display more detail on something by clicking on some reduced representation of that object is the real legacy of hypertext research. How that behavior is resolved by the systems involved is not important (and often unknowable) to the user; what matters is the experience.