innovation in metadata design, implementation & best practices

Topic: DCMI Abstract Model
Modified: 2004-03-22 09:41, Monday
Maintainer: Tom Baker
Latest version:
See also:

Shepherd: Andy


Andy has revised the draft model:

We need to discuss:

-- aspects of this model that relate to Usage Board 
   practice, such as the notion of Application Profile;

-- use of dcterms:URI as the encoding scheme for those parts of
   encoded metadata that express a "value URI" as opposed to 
   a "value string" (see email below);

-- the relation of this document to other "foundational" 
   documents maintained by DCMI, especially, which is 
   included in the main packet (see discussion below).

Notes from Tom on the relation between the DCMI Abstract Model (DAM)
and the existing DCMI Grammatical Principles (DGP) at

I see two slightly different needs:

1) The need to present this new model to the general public
   in a self-contained and readable form.

2) The need for DCMI to have a set of URI-identified documents
   that makes optimal sense from the standpoints of citation
   and maintenance.

Need #1 is (almost) met by the existing draft, which
contains almost all of the elements one would need for a
self-contained presentation to a broad public. For this,
in my opinion, it still needs a narrative introduction -- one
which wraps the model in some historical context and briefly
states how the present document relates to earlier documents.
With such a context, the article would basically be ready for
D-Lib Magazine (let's call this "DLIB-DAM" as opposed to a
"DCMI-DAM" to be maintained as a DCMI document).

In terms of Need #2, however, and with respect to
other DCMI-maintained documents, the DAM "package"
(model plus Appendixes) does not seem ideal:

-- The existing two-page "DCMI Grammatical Principles"
   ( is almost
   completely subsumed in Section 8 (Terminology) of DAM.

   The only things missing from DAM with respect to DGP are:
   -- a two-sentence scope statement that explicitly gets the 
      definition of "term" from the DCMI Namespace Policy;
   -- a one-sentence definition of "appropriate value";
   -- a three-sentence discussion of Vocabulary Terms (with
      the recent addition of "Broader Than" and "Narrower 
   -- a one-sentence definition of "Application Profile" 
      (along the lines of the one I suggested for DAM);
   -- a short footnote acknowledging a historical shift 
      towards seeing Element Refinements as Properties.

   Having to maintain DAM and DGP in sync creates potential
   error and confusion (and extra work). The solution could be
   either to delete DAM's Section 8 ("Terminology") and replace
   it with a pointer to DGP -- or vice versa. I am inclined
   to the latter, i.e. to delete the DGP document (redirecting
   its URI) and add some (but not all!) of the missing bits
   to DAM. Then the terminology section (Section 8) could
   perhaps be moved up to the front of DCMI-DAM -- boring
   for a D-Lib article but good practice for a specification.

-- Rather than leave the comparisons of RDF, XML, and XHTML
   encoding specs against the model in the DCMI-DAM document
   itself, I might sense either:

   1) to fold each of the comparisons into the Specification
      in question; 
   2) to break out each comparison as a separate document
      with a citable URI and (in the longer term) subject to a
      separate processe of revision and approval; or 
   3) to bundle all of the comparisons into a single,
      citable document (but not as an Appendix to the model

   On the other hand, the appendix on RDF as it is currently
   written can help people who already know RDF understand
   the Abstract Model itself, so perhaps it should remain
   in the Appendix even if yet-to-be-written comparisons for
   DCMI's two RDF-related Specifications reside elsewhere.

   I do not have a strong opinion about these options.

-- The note about structured values is on one level
   an abstract typology, but it is also framed as a response
   to the state of discussion as of the time of writing.
   One could perhaps split it out as a separate document with
   a separately citable URI, though it is only one page long
   so I do not feel strongly about this.

Note that following the suggestions would have the effect of
making DCMI-DAM document shorter (just five pages) and turning
the existing two-page "DCMI Grammatical Principles" into a
redirect to DCMI-DAM. 

Date: Wed, 25 Feb 2004 10:49:27 +0000
From: Andy Powell <a.powell@UKOLN.AC.UK>

When a vocabulary term gets a URI assigned to it, it changes
from being a 'value string' to being a 'value URI' and the
encoding needs to change to reflect that. So for example in
XML, instead of

<dc:subject xsi:type="dcterms:DDC">Internet</dc:subject>


<dc:subject xsi:type="dcterms:DDC">004.678</dc:subject>

depending on your preference for numbers or words, the encoding would
change to something more like

<dc:subject xsi:type="dcterms:URI">info:ddc/22/eng//004.678</dc:subject>

As usual, the RDF/XML encoding handles this rather better.

I'll write up something longer about this before the meeting. This is
just to flag up the issue.