Re: There Are No Metadocuments

At 8:38 PM 2/10/97, Terry Allen wrote:
>Strange.  I can publish a print book, which binds together in
>what everyone recognizes as a book font, size, kerning, paper
>type, thumbtabs, page numbers, line numbers, verse numbers,
>and footnote markers that cause the eyes of literate viewers
>to flicker involuntarily to the bottom of the page (behavior).

>Yet when I bind all this together (with the text) in a document,
>somehow this becomes a metadocument.  If I reprinted the book
>in another font, font size, etc., I'd have another book,
>perhaps "The Large Type Document for Presbyopters," which
>I would call attention to as different from "The Document,"
>although for some purposes it is "the same".

For the kinds of purposes that SGML (and XML) propose to support -- i.e.
re-use they _are_ the same. That is why some of us are so vociferous about
separating processing from data content (and insistent as Len says, that we
have only a bag of (structured) values, behavior and rendering free).

Anytime we break that distinction, we make some form a reuse more difficult
or impossible (depending on how central to the document markup the
formatting decisions become).

Now, if these reuse issues are not appropriate in a WWW context, then we
should not be doing XML at all, but working on adding more kinds of
behavior tag to HTML so that it can finish its metamorhphosis into
"Postscript for the Web".

This is all baby SGML, folks. It's the _point of the whole thing_. So let's
finish it, but stop dragging over the same issues. I don't mind Postscript
for the web, but if that's our goal then we don't need generic tagging to
get there.

>This group must eventually determine how to associate all the
>information that comprises the "metadocument".  It must also
>determine how to associate the parts of a "document".  Why
>should these be separate tasks?

Who cares. If we want the rendering/processing independent part of the
document to retain those properties it must be separated from the
rendering/processing instructions. So we need a mechanism to "glue" them
together.

>And where is the description of the overall architecture?]

                            _________________
                           |                 |
[document]---------------> |  User viewing   |
                           |   Engine        |
[link-action spec]-------> |     or          |
                           |                 |
[text layout spec]-------> | Web Walker      |
                      ^    |     or          |
                      |    | Something else  |
                      |     _________________
                      +------------------------User choice and publisher choice
                                               may affect/replace these bindings
                                               and the contents of the specs.
                                               (therefore there is a separate
                                                binding entity that controls the
                                               existence of those arrows)
                                               [ secure systems will need to
                                                 limit user choice, but the
                                                 architecturally, all processing
                                                 agents are equivalent ]


_________________________________________
David Durand              dgd@cs.bu.edu  \  david@dynamicDiagrams.com
Boston University Computer Science        \  Sr. Analyst
http://www.cs.bu.edu/students/grads/dgd/   \  Dynamic Diagrams
--------------------------------------------\  http://dynamicDiagrams.com/
MAPA: mapping for the WWW                    \__________________________

Received on Tuesday, 11 February 1997 13:26:15 UTC