>>There will be great pressure from the publishing industry, where I come
>>from, to focus primarily on the richness of the markup of individual
>>documents. For the most part, those people should be ignored! They don't
>
> Define markup, and I'll decide whether to agree or not.
That's a good question. I'm drawing a distinction between markup that is
structural (within a document) and markup that is navigational (among
documents), except that if I'm not careful, I'll be trapped in a tautology.
Clearly, any kind of markup could potentially be interpreted for
navigational purposes.
I guess the difference is the explicit purpose of the markup. For example,
tagging the words, "Bill Gates," with the information that he is chairman
of Microsoft seems to me to be explicitly for navigational purposes. It
doesn't define the structure of the document in which is contained; it can
be used to define the relationship of the document to other documents.
>One of the reasons why I have been asking for arbitrary SGML, and
>arbitrary stylesheets is to aid navigation. An SGML document is (or
>should be) basically a tree. One can use such structure for
>sub-document navigtion, which might be important for very large
>documents. The thing with stylesheets is perhaps not so much for
>navigation, but with a generic stylesheet mechanism, parts of a
>document can be formatted to best help people understand the
>information. Such formatting might include removing large parts of
>irrelevant data (for example, if you had a play, you might remove
>everything *except* the names of the speakers, and when you clicked on
>the speaker name, a box containing the speakers lines might appear).
>This might seem like formatting, but it *is* related to navigation and
>understanding via visual, and perhaps aural clues.
>
>I agree that net navigation is important. I also believe sub-document
>navigation is important. Let's try to support both.
I've never really made the distinction between net navigation and
sub-document navigation. That's a matter of ownership and physical
location, primarily, I think.
I'm not sure I understand about the tree structure. Surely you aren't
implying that *all* SGML documents should have a hierarchical structure,
which is how I interpret "tree." Are you?
On the other hand, I have created documents like the one you describe,
which clearly are a great benefit of hypertext. I love documents and
sub-documents, as you say, that expand and collapse according to complex
rules (complex to the programmer, that is).
Nick
Multimedia Computing Corp.
Campbell, California
----------------------------------------------------------
"We are surrounded by insurmountable opportunity." -- Pogo