Streams, protocols, documents and fragments

Michael.Orr at Michael.Orr at
Fri Feb 26 22:19:31 GMT 1999

> -----Original Message-----
> From: keshlam at [mailto:keshlam at]
> Subject: RE: Streams, protocols, documents and fragments
> Personal opinion: The right way out of the "never-ending document" problem
> is to declare that the stream is a stream of transaction documents, NOT a
> single huge document in its own right.

Exactly. In many cases -- I suspect the vast majority -- the XML DTD or
schema, parse bite size, and transaction considerations all map together
very effectively at a relatively fine granularity. 

The remaining questions, at the level of the containing interchange
stream(s), can be approached as protocol design and separated by
layering from the document considerations. Using document modeling to
describe the stream would simultaneously overkill the structure side and
fail to engage with protocol state issues. 

This is not to deny the requirement for document scalability -- there
are very good reasons to make sure that XML tools are prepared to cope
with huge documents. But: a huge document is not a good implementation
for a transaction stream. 


Michael Orr, CTO, VP R&D
Design Intelligence Inc, Seattle WA USA
pager:888-688-4609  fax:206-343-7750

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at
Archived as: and on CD-ROM/ISBN 981-02-3594-1
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