Print

Print


Pete,

Thank you for your detailed reply to my question and for the
clarification.

I think that when constructing the profile the group has to be doing

(a) creating a set of terms, using the experience of the RSLP
Schema as a starting point?

rather than

(b) referencing/reusing the existing terms in the RSLP element set?

simply because RSLP does not have any formal recognition or
namespace. But this seems a pity when RSLP CD is already
defined and in use.

Are there other standards or generally used schemas, or
developing schemas, for collection descriptions, which the group
could use? Maybe you have already looked into this. Why not use
EAD? - is it thought to be over-complicated?

You note that RSLP CD is already emerging as a 'de facto'
standard within certain communities. This is also my impression.
But I suspect this is only within the UK. I notice that most of the
replies you've had to this list have been from the UK. I wonder
whether different schemas are being used elsewhere.

I support your intention to develop a DC-CD schema based on
RSLP, but think also that some of these questions need answering
within the international DC community. Maybe we could have some
fruitful discussion at DC2002?

Best wishes,
        Ann



Date sent:              Fri, 26 Jul 2002 12:00:59 +0100
Send reply to:          DCMI Collection Description Group <[log in to unmask]>
From:                   Pete Johnston <[log in to unmask]>
Subject:                [DC-COLLECTIONS] Application profiles and element reuse
To:                     [log in to unmask]

> [I've broken out this part of Ann's message into a separate thread, as
> I think it raises some important points about the way we approach the
> construction of an application profile. I talked to Andy about this
> earlier and I hope this is a reasonable representation of our
> discussion - but Andy may wish to add his comments]
>
> > I do have some concern about using RSLP for a DC-endorsed
> > application profile. It seems there is an intention to extend
> > DC to include the CLD elements? But RSLP is not a recognised
> > standard, so does what is planned fit in with DCMI strategy?
> > If this application profile were agreed would the CLD
> > elements then become part of DC?  Is it a means of achieving
> > some recognition for RSLP? I'm not necessarily disagreeing
> > with this approach but would like some clarification.
>
> First I guess we should be clear to everyone on the list that Andy and
> I, the co-chairs of the WG, are both based at UKOLN and have both been
> involved with work related to the RSLP CD Schema. Andy did most of the
> work on the Research Support Libraries Programme (RSLP) Collection
> Description project, based at UKOLN a couple of years ago, which
> developed the RSLP CD Schema. Part of my time is spent on the
> Collection Description Focus post, which is partly funded by RSLP, and
> part of that work involves supporting implementers of the RSLP CD
> Schema.
>
> A DC-based application profile designed for a specific domain or
> resource type, like the one this WG is seeking to develop, will
> typically include some terms which are specific to that domain or
> resource type, and which are not presently associated with the DCMI
> http://purl.org/dc/elements/1.1/ or http://purl.org/dc/terms/
> namespaces. Those new terms might be:
>
> (a) created by the developers of the profile for their specific
> requirements, and associated with a new namespace (which in this case
> would be a a namespace associated with the DCMI), and/or (b) drawn
> from other existing element sets associated with namespaces
> owned/managed by other bodies.
>
> With the non-DC terms listed at
>
> http://homes.ukoln.ac.uk/~lispj/dc-cd/rslpcd.html
>
> I was working on the basis that essentially this WG was doing (a) i.e.
> _creating_ terms (let's call them dc-cd: terms, for now for the sake
> of argument) rather than strictly _reusing_ the existing terms from
> the RSLP CD element set. In that process of creating terms we are
> using the terms of the RSLP CD element set as a starting point for
> discussion, but these dc-cd: terms would be separate terms from the
> RSLP CD terms (i.e. in an RDF implementation would be identified by
> different URIs). On that basis we could choose that the semantics of,
> say, dc-cd:hasAssociation was different from the RSLP CD term
> hasAssociation, and we could use different names for the elements even
> if the semantics were similar, and so on. (We could also make
> statements about semantic relationships between terms from the two
> element sets, e.g. using rdfs:subPropertyOf or daml:equivalentTo etc)
>
> Working on this basis, it seems to me the activity of the DC CD WG
> says nothing specifically about the status of the RSLP CD Schema,
> other than "this is something out there which provided a good starting
> point because it seemed to work for them", and any DCMI Usage Board
> endorsement (or otherwise!) of an application profile produced by this
> WG would not say anything about the RSLP CD Schema.
>
> On the question of whether we propose all the new profile-specific
> terms to Usage Board for association with, say, the
> http://purl.org/dc/terms/ namespace, I understand that the DCMI Usage
> Board is at present working on a statement to clarify some of these
> issues, but my understanding is that, yes, that is the approach that
> is being taken with other application profiles proposed by DCMI WGs.
> The decision about whether they were accepted or not would rest with
> the Usage Board.
>
> If, on the other hand, we are doing (b) i.e. _reusing_ terms from the
> RSLP CD element set, then there are some other considerations. In
> practical terms I think that implies more constraints on the work of
> this WG, because this WG can not alter the semantics of terms defined
> by the RSLP CD Schema (other than perhaps to say we are using those
> terms in a particular way). I guess there is also something of an
> implicit endorsement of the RSLP CD Schema - which, as you say, is not
> a "standard" in any formal sense (though it may be emerging as a de
> facto standard within certain communities). In this case, the terms
> would not be submitted for inclusion in the http://purl.org/dc/terms/
> namespace, and they would not be "DCMI terms".
>
> There may be some other subsidiary considerations here which might
> need to be addressed on the RSLP CD Schema side of things e.g. the
> terms in the RSLP CD element set are associated with the namespace
> name/URI http://www.ukoln.ac.uk/metadata/rslp/1.0/ and there's
> probably an argument that it might be better to choose a URI which
> explicitly carries some promise of persistence e.g. a PURL. Strictly
> speaking that is probably an issue for Andy and myself wering our RSLP
> Schema hats!
>
> But I think we do need to address the fundamental question of whether
> this group is constructing a profile by
>
> (a) creating a set of terms, using the experience of the RSLP Schema
> as a starting point?
>
> Or
>
> (b) referencing/reusing the existing terms in the RSLP element set?
>
> Cheers
>
> Pete
>


--------------------------------------------------------------------------
Mrs. Ann Apps. Senior Analyst - Research & Development, MIMAS,
     University of Manchester, Oxford Road, Manchester, M13 9PL, UK
Tel: +44 (0) 161 275 6039    Fax: +44 (0) 0161 275 6040
Email: [log in to unmask]  WWW: http://epub.mimas.ac.uk/ann.html
--------------------------------------------------------------------------