I was an avid user of Doug's NLS during the period of ARPANET host protocol and early Internet protocol development. We used Texas Instruments Silent 700 model terminal with building in acoustic couplers and thermal paper. Incredibly, NLS worked well for that primitive mode of operation. The ability to perform massive re-organization of text, to view and manipulate using high levels of abstractions and to link documents together made NLS a very powerful tool indeed.
I would place collaborative environments at a very high level in Doug's universe - allowing people to register and share information in the form of documents and the ease of moving back and forth among them. An evolution of this in the form of Google Docs that can be edited concurrently by multiple parties is enormously useful, especially if a group is using voice/video teleconferencing to view and edit the object concurrently. Having multiple documents open at the same time for group use is also very useful.
The infrastructure needed to achieve these effects is significant and this leads to yet another significant question: how long can these objects be maintained in useful form? what happens to them when the significant infrastructure that allows them to be accessed so freely is no longer available? What aspects of the collection of documents can be preserved in perpetuity? What would permanent links to these digital objects look like and how would they be resolved? What metadata should be preserved with each digital document (object?)? What access controls should be available? How should individuals identify themselves to establish their bona fides to access the objects? What can each individual do to/with each object? What is the resolution of reference to each digital object? A word? A paragraph? a letter? a pixel? More generally, the entire panoply of questions arising in the context of preservation of access to digital content are readily invoked here.