Content or Metadata?

Robin Cover robin at
Thu Dec 2 23:02:14 GMT 1999


In some OO theory, I think it's believed favorable to create
distinct attributes for things that are ordered (since 
[Boyce-] Codd believed that attributes are intrinsically
unordered): this, for Mike Spreitzer's example of
"list of authors": firstAuthor, secondAuthor, thirdAuthor,
etc.  Well, suppose there are in fact three groups of
authors, with different principles of sub-ordering, which
are masked in the typical presentation... it may then be
more economical (80:20 rule, which I detest) to say that
we allow an attribute value which is an orderable list
of (sub-)tokens.  I have seen -- indeed, documented -- some
works which enumerated over 30 authors for the piece.
Volumes/analytical works from the French academies.
(And why not?  Only the aesthetics of print books and the
supposed cost of printer's ink have lead to style rules that
say "truncate with 'etc' after N authors...".)  In such
cases: I suspect the order (-edness, -ability) has nothing
to do with whether the factoids are (meta-)data or not.

Nothing is simple, despite what could appear to be
incontrovertible facts.



On Thu, 2 Dec 1999, Tim Bray wrote:

> At 11:41 AM 12/2/99 PST, Mike Spreitzer wrote:
> >What about the list of authors of a scholarly paper?  Isn't that metadata for which order
> >matters?
> Yep, in fact that's the one use-case that kept coming up during the early 
> stage of RDF design.  Here's another one for free: content models.  But
> the notion that there is some ordering on a document's author, title, 
> and date-of-publication is surprising and unnatural. -T.

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 unsubscribe, mailto:majordomo at the following message;
unsubscribe 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