> The three layer model he's proposing doesn't seem very different from
> Durusau's Just In Time Trees combined with a style sheet language and I
> wouldn't call it a bad idea.
Or Attributed Range Algebra... in fact one of the uses of ARA was to
enable something like that. One of the demo's I have here is a database
that takes in XML, LaTeX (restricted form) and email messages, and
creates a unified markup database/index from it. Anything can be pulled
put as XML, treated as XML using a DOM (single implementation for all
markup), markup can be suppressed or generated on the fly, etc. The
difference is in treating markup as an something interpreted/inferred
rather than something truly intrinsic to the data.
I think there is a great deal of power in the model, and remember that
Ted Nelson's perspective is usually tinted by the desire for persistent
hypermedia linking infrastructure.
