HC and hypertext communities aren't talking ... ?

It seems like there should be a lot more communication between the Human-Computer Interaction community and the hypertext community. Right now, most of the work between those two areas seems to be largely aesthetic: how do you design a web page so that people can navigate it, and what features do you put into your site to allow readers to do what they're trying to (or, perhaps, find things they weren't expecting)?

But there's so much more that could be going on! People don't only Interact with computers, they think with computers. One reason that hypertext is such a powerful tool is because it offers the potential for people to work with their information in ways that are more loyal to the way they think. Good hypertext tools usually solicit pages of raves by new users who are amazed to finally find a tool that facilitates they way they really work.

So who is:

  1. applying the research on hypertext to the ways that people actually read it?

  2. taking the research on how people browse websites (probably the largest dataset on how people read hypertext)

  3. applying that to the ways that we design hypertexts to be readable?

  4. taking the knowledge of the way that people work with computers and applying it to hypertext tools?

This stuff must be out there, but it's hard to find examples of people combining the study of how people interact with computers with the study of how people work with hypertext. "Hypertext" ought to be smack in the middle of people-and-computers. I think that Eastgate Systems has hit #4 with Tinderbox, but the prohibitive learning curve indicates that they've got a ways to go in terms of UI. I think that Dynamic Diagrams is a good example of some of the other points.