Bob Stayton (
Wed, 11 Aug 93 10:25:38 PDT

> From: Lou Burnard <>
> It seems to me that the LINKs that people are currently talking about
> are an artifice which HTML has to bolt on as a direct consequence of not
> providing proper document structuring capabilities. (hastily dons
> asbestos overalls)
> In a *real* SGML world, your document structure would be exactly
> mirrored by the SGML structure. In HTML-ese, your H1 elements would nest
> directly within the document, H2 elements within H1s and so on down.

This came up during the discussion of HTML+ at WWWWW.
I agree that SGML DTDs usually nest section containers to
indicate the structure of the document. I have a concern
that this conflicts with the chunking of information into
small files for usable performance at delivery time. I
have a 1000 page System Administrator's Guide that I want
to put on line. I cannot put that into one document file,
because the time to transmit that whole thing over the net
for one HREF link is not good for the network or the user.
And the user had better have a big cache for their
second HREF |^).

I need to break this large hierarchical document into
lots of smaller files, divided mostly along section
boundaries. Then an HREF delivers a small chunk
quickly. But if the DTD specifies nested sections,
can each file be valid SGML? Here are two scenarios
(using hypothetical <S> section containers).

Scenario 1: Scenario 2:
<S3> <S1>
<TITLE>Printer setup</TITLE> <S2>
blah <S3>
</S3> <TITLE>Printer setup</TITLE>
</BODY> blah

Scenario 1 is an <S3> element without its <S1> and <S2>
containers, and most hierarchical DTDs would not consider
this document fragment parseable. Of course a DTD could
be written to allow the BODY to start with any section
level. But haven't you lost the hierarchy for this
document then?

Scenario 2 has phony placeholding containers that
actually occur in other files. Although it is parseable, it
is clearly an abuse of SGML.

In either case, I still would like to
provide information to connect the <S3> section to its
parent, children, and siblings. The very act of
breaking up the large document means I have to
provide some mechanism to indicate those connections,
either with the LINK mechanism in HTML or as some
IDREF attributes in the <S> tags. There should
be enough information to reassemble the
hierarchy starting from any point in the hierarchy.

Bob Stayton 425 Encinal Street
Technical Publications Santa Cruz, CA 95060
The Santa Cruz Operation, Inc. (408) 425-7222