> I think if you look at RDF as markup, you miss the
> point of RDF.

It might be useful at this point to make a distinction between
metadata tagsets, and data tagsets... this is a distinction that I
have rarely been fully happy with, but as time goes on I am becoming
more enamoured with it:-

Tagsets for: UI (User Interface) oriented structural textual rendering
(e.g. Docbook, HTML, MenuML, OEB), specialized rendering (e.g. MathML,
SVG - Scalable Vector Graphics, MusicML, SMIL - Synchronized
Multimedia Integration Language), or any generic data storage format.
An informal definition is 'anything for which the question "is there a
textual equivalent of all rich media data bits" makes sense'.
When the content being marked up is metadata. Examples: For expressing
data processing (e.g. XSL - Extensible Style Language), RDF (Resource
Description Framework), Schema languages, etc.

Wed, 21 Mar 2001 08:44:37 UTC

Because RDF is intended for machine processing, rather than delivering
it to a client, we get mixed up when it is used in a generic data
tagset, i.e. XHTML. In other words, it's the old "embredding XML into
XHTML" problem - it doesn't make all that much sense. What does make
sense is partitioning off applications of RDF, and embedding *them*. I
agree that limiting ourselves to just Dublin Core would be silly, but
it's a good a place to start as any. Next on the list would probably
be RDFS, so that people can create RDF Schemas, human annotated.

Kindest Regards,
Sean B. Palmer
@prefix : <> .
:Sean :hasHomepage <> .

Received on Monday, 16 April 2001 15:28:01 UTC