Compound Documents - necessary for success?

Marcus Carr mrc at
Thu Jan 28 04:39:12 GMT 1999

Roger L. Costello wrote:

> Could we discuss the pros and cons of compound documents?  First let me
> define what I mean by "compound document":
> compound-document ::= (compound-document | valid-document)
> valid-document ::= <a document that conforms to a schema>

I would love to discuss compound documents - having grappled with the same
problems in SGML implementations with very little success, I'm curious as to how
it will work with XML.

> In words, a compound document is a "document of documents", where each
> document conforms to a schema; i.e, a nested document conforms to a schema
> as well as does its parent document.  I will use the term composition and
> compound document interchangeably.

One problem that I've seen in the past is the issue of multiple levels of
nesting, often with the same DTD being used. In legal case reports for example,
it's not uncommon to see the judge include a fragment from another case report
that includes a fragment from a third report. (In a legal system based on
precedence, this is entirely reasonable.)

What I don't understand is whether these fragments would use the same namespace
as the parent document does. They do use the same DTD, but they're part of a
different document. If they do use the same namespace, does this not introduce
difficulty in the creation of an XSL stylesheet? After all, you may need to
format the elements differently based on their ancestry, but also perhaps
indenting the text further to reflect the fact that these documents are nested.
It's not realistic to suggest that I just allow the documents to exist in the
DTD, eliminating the need for calling this a compound document - the data
logically lives in other structured documents and besides, it would make the DTD
so loose as to be useless for any other purpose.

Also, what would prevent someone from putting together a compound document
constructed entirely of valid fragments? To the naked eye it would look like an
invalid document, but the fact that the fragments themselves were valid and
existed in a location that allowed them would mean that the document was parsable
without error. I'm not suggesting this would be a common way of producing
documents, but nor do I think it's a good idea that you could conceivably create
a class of documents between well-formed and valid.

Along similar lines, how would I include, say, two paragraphs of a judge's
speech? Does a compound document imply a single doctype element? Would I have to
wrap the two paragraphs in some artificial element just to make it a complete
document? What would direct a user to this solution over adding each paragraph as
a separate document?

It seems that much effort has been spent in accomodating compound documents, but
to the untrained eye, it seems that the support has been built prior to the
problem being fully understood (by me, at least...:-). It appears that the 'XML
Fragment Interchange Requirements' is being designed to address the management of
fragments - is it anticipated that this will complement or obviate the need for
namespaces in compound documents?


Marcus Carr                      email:  mrc at
Allette Systems (Australia)      www:
"Everything should be made as simple as possible, but not simpler."
       - Einstein

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at
Archived as:
To (un)subscribe, mailto:majordomo at the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at

More information about the Xml-dev mailing list