I will be out of the office Monday 4/21 - Friday 4/25. If you need
immediate assistence please contact Layton Elliott at 503.494.1656.
-dls-
>>> "[log in to unmask]" 04/23/08 03:46 >>>
Hi Tom, Mikael,
> Public Comment is being held on the Working Draft
>
> "Description Set Profiles: A constraint language for Dublin
> Core Application Profiles" [1]
>
> (and accompanying XML schema for the XML representation [2]) through
>
> Monday, 28 April 2008
>
> As described in the "Singapore Framework for Dublin Core
> Application Profiles" [3], this specification offers a simple
> constraint language for Dublin Core metadata on the basis of
> the DCMI Abstract Model [4]. Interested members of the public
> are invited to post comments to this mailing list [5],
> including "[Public Comment]" in the subject line.
Thanks. This is looking great.
A few very minor comments on the syntaxes:
(i) the Description and Statement Templates both carry min/max
occurrences constraints, which are represented by XML attributes called
minOccurs and maxOccurs, and that dovetails nicely with the convention
used in W3C XML Schema.
There is also a second category of constraints, dealing with (I think)
components which can occur at most once, but for which we want to
indicate whether they are "mandatory" / "optional" / "disallowed", and
these are represented using XML elements with the names
LanguageOccurrence, SyntaxEncodingSchemeOccurrence, ValueURIOccurrence
or VocabularyEncodingSchemeOccurrence, with element content one of those
three strings.
I think the analogous role in W3C XML Schema is played by the "use"
attribute which takes corresponding values
"required"/"optional"/"prohibited". While I recognise there's absolutely
no obligation at all to align with W3C XML Schema, I wondered whether it
was worth adopting that convention, just so that people working across
the two don't have to keep making a mental adjustment, i.e. using
something like
<LanguageUse>required</LanguageUse>
intead of
<LanguageOccurrence>mandatory</LanguageOccurrence>
(ii) in the RDF syntax the analogues of the minOccurs and maxOccurs XML
attributes are properties mapped to the QNames dsp:minOccur and
dsp:maxOccur, and I wondered whether for the sake of consistency it was
worth changing the property URIs so that the latter QNames were
dsp:minOccurs and dsp:maxOccurs (i.e. with a final "-s") instead.
(iii) in the skeleton examples, both in the XML case and the RDF case,
there's a literal list constraint instance with two options, and each
option is shown with both lang and SES attributes (in the XML case) and
with both xml:lang and rdf:datatype attributes in the RDF/XML case. I
think - for one specific Literal Option within a list - the
language/datatype attributes will be mutually exclusive, right?
I do have another question that I've been fretting about, which I think
is really about the Singapore Framework and the relationship of DSPs to
DCAPs, but I'll put that in a separate message.
Pete
---
Pete Johnston
Technical Researcher, Eduserv Foundation
Web: http://www.eduserv.org.uk/foundation/people/petejohnston/
Weblog: http://efoundations.typepad.com/efoundations/
Email: [log in to unmask]
Tel: +44 (0)1225 474323
|