I've created a new draft of the DC CD AP:
http://www.ukoln.ac.uk/metadata/dcmi/collection-ap-summary/
http://www.ukoln.ac.uk/metadata/dcmi/collection-application-profile/
Bearing in mind that the Usage Board will be looking at this perhaps
less from the perspective of whether the DC CD AP meets our "functional
requirements", and rather more from the perspective of whether it is
"well-formed" with respect to the DCMI Abstract Model (and whether it
makes reasonable use of individual terms from the DC vocabularies), the
main changes I've made are - as suggested at [1] - to try to bring it
into alignment with the terminology and concepts of the DCAM, by
1. providing a bit more introductory commentary on what the DC CD AP is;
2. tweaking the text, mostly of "Comments", here and there to refer to
values, value strings, statements etc as appropriate;
3. reorgansing the tables a bit and for each property, for each
vocabulary encoding scheme recommended for that property, adding a table
which indicates whether a value URI, value string or rich representation
is required/optional.
I've assumed that for every value, a related description _may_ be
provided - so you might have a description of a subject term or a
location or a collector or a service or whatever, but this DCAP doesn't
tell you anything about what terms might be used in those descriptions.
I hope (3) goes some way to addressing the questions I was getting along
the lines of "Do I use a URI or a name/title here?" a bit better than my
stock response of "Well, you can use either but you need to be sure
which you want...". I've probably made some fairly arbitrary decisions
along the way here, but I've tried to base them on what I've seen of
existing practice, particularly in e.g. the JISC IESR project.
So e.g. for properties like dc:title, dcterms:abstract, a "value string"
is required but a "value URI" is optional; and for properties like
cld:logo where the value is an image, a "value URI" is required, but a
"value string" is optional. I was strongly inclined to mandate the use
of a "value URI" for the relations between collections, but given that
we stopped short of saying that collections must have URIs assigned to
them (I think that was what we settled on? The use of dc:identifier with
a URI is "optional but recommended" in the current version?), I've left
it as optional. I made the decision that "rich representations" of
values are not supported.
We can discuss some of this and if necessary revise the detail of these
options later - I suggest it might be a good topic for the f2f meeting
of the WG in Madrid - , but I wanted to get a first cut at this in place
in the version that the Usage Board review, as I think it illustrates
some issues raised by the DCAM which have not been addressed by previous
efforts at modelling/representing DC application profiles.
In terms of the "content" of the DC CD AP, the main changes I made were:
- Format of Items/Collection: I've removed the dc:format property,
pending resolution of the question of how to represent the fact that a
collection has items of a specified format
- Date ranges: The DC Date WG hasn't yet made a recommendation on the
issue of how to represent "open date ranges" (which neither W3CDTF or
ISO8601 support). So I've opted to reference an encoding scheme based on
the date format used by the Australian Recordkeeping Metadata Schema
(RKMS) [2]. I've created a (currently temporary) URI for this (see [3] -
though at present I'm not suggesting we propose that scheme to UB - date
schemes is a job for the DC Date WG) - as I couldn't see one defined by
RKMS (but if anyone can point me at an existing URI, I'll use that).
(The other option would be to define a new date format, which I also
made a start on. See [4])
I'm happy to make minor changes to this version of the DC CD AP before
passing it to the Usage Board (e.g. I haven't changed the logo
definition but if we can agree wording for that - which I get a sense we
are close to doing - then I'll include it in a new version), but I'd
like to post a version to the Usage Board first thing on Monday, so any
comments need to be made by the end of this Friday (19 August).
I have very little time (well, none, really) left to work on this this
week, so I'm afraid any substantial changes will have to wait (unless
there are any glaring errors) ;-)
There is one issue I'm slightly uncomfortable about but I don't know
what to do about - I'll post a separate message on that, probably
tomorrow.
Cheers
Pete
[1]
http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0504&L=dc-collections&P
=1125
[2] RKMS Schemes
http://www.sims.monash.edu.au/research/rcrg/research/spirt/deliver/schem
es.html
[3] http://www.ukoln.ac.uk/metadata/dcmi/collection-RKMS-ISO8601/
[4] http://www.ukoln.ac.uk/metadata/dcmi/date-dccd-odrf/
-------
Pete Johnston
Research Officer (Interoperability)
UKOLN, University of Bath, Bath BA2 7AY, UK
tel: +44 (0)1225 383619 fax: +44 (0)1225 386838
mailto:[log in to unmask]
http://www.ukoln.ac.uk/ukoln/staff/p.johnston/
|