Commentary on DCD

[John Cowan]

Here are a few tentative comments on the DCD proposal.  I am
writing from my experience developing XSchema, but I am not writing
*for* the XSchema team.

1. The Description and Value properties seem to extend the RDF
syntax in ways not warranted by the current RDF draft (more bluntly:
"are erroneous RDF").  The value of these properties is evidently
a Resource which is encapsulated within the property element.
An example appears in clause 3.4.2 of DCD.

However, the RDF syntax in the current draft (1998-07-20) supports only
three kinds of content for property elements:  #PCDATA representing
the string value of the property, a container element as value
of the property, and an RDF:Description element serving as a proxy
object.  There are no provisions for embedding arbitrary mixed
content and calling that the value of the property: data can't
go inside metadata, roughly speaking.

To make Content and Value conform to the current draft, they would
need to be pulled out of the RDF part of the document altogether,
tagged with an ID attribute, and referred to from within the RDF
using <DCD:Value about="#the-id"> or the like.

Clause 2.1.1 is silent on this extension to RDF.


2. Based on XSchema discussions, I believe that more suitable values
for the Root property would be "Recommended", "Unlikely", and
"Forbidden" or the like.  Distinguishing between "Recommended"
and "Unlikely" assists authoring tools in limiting normal use
of document roots without constraining abnormal but still
permissible uses, thus furthering design principle 3.


3.  The DCD draft is totally silent about notations, in violation
of its own design principle 1 (clause 1.2)  Although notations are not
part of namespaces, they are still an important factor in XML.

Perhaps as a consequence, the datatype "notation" in clause 4.1
fails to provide for the list of legal notations that the XML
attribute-type model requires: there are no XML attributes of
type "notation" simpliciter.


4. The list of primitive datatypes is still too long and cluttered.
Id, idref, and idrefs are really NAME and NAMES with the separate
id-roles property.  Fixed-point values should be represented with a
denominator property rather than the less general (decimal) scale.
General rational numbers need support (353/113, e.g.).  The
various forms of date-time are clutter: there should be only one
kind of timestamp, useless SQL specifications notwithstanding.
All the i2, i4, etc. etc. are just values of max and min.
There is no way to represent the difference between exact and
inexact values.  (Yes, I know I'm making work for everyone.)


-- 
John Cowan	http://www.ccil.org/~cowan		cowan at ccil.org
	You tollerday donsk?  N.  You tolkatiff scowegian?  Nn.
	You spigotty anglease?  Nnn.  You phonio saxo?  Nnnn.
		Clear all so!  'Tis a Jute.... (Finnegans Wake 16.5)

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