DCMI Basic Page Template

Date: Mon, 17 Sep 2001 22:13:30 +0100
Reply-To: A mailing list for the Dublin Core Metadata Initiative's Usage
Working Grou <DC-USAGE@JISCMAIL.AC.UK>
Sender: A mailing list for the Dublin Core Metadata Initiative's Usage
Working Grou <DC-USAGE@JISCMAIL.AC.UK>
From: Andy Powell <a.powell@UKOLN.AC.UK>
Subject: Re: Encoding schemes for dc:identifier
In-Reply-To: <Pine.GSO.4.10.10109171751470.23399-100000@twister.gmd.de>
Content-Type: TEXT/PLAIN; charset=US-ASCII
On Mon, 17 Sep 2001, Thomas Baker wrote:

> Dear all,
>
> In the context of our fast-track discussion in Tokyo, could someone
> please volunteer to discuss the issue below (encoding schemes for
> dc:identifier)?
>
> Note that in follow-up discussion, Stu felt that DCMI should register
> such schemes as they come to our attention, whereas Thomas Krichel and
> Ray Denenberg referred to the RFC2288 proposal to encode conventional
> bibliographic identifiers as URNs [1]. The status of such proposals is
> apparently tracked at [2].
>
> It would be very useful if someone could summarize this issue for our
> consideration and suggest some guidelines we could follow.

The list of registered URI schemes is at:

http://www.iana.org/assignments/uri-schemes

The list of URN NIDs is at

http://www.uri.net/urn-nid-status.html

DOI doesn't appear in either. ISBN, ISSN and SICI appear in the URN NID
list.

doi:10.1000/182 is a perfectly good URI. It just isn't registered.

It is perfectly valid to encode

<dc:identifier scheme="URI">doi:10.1000/182</dc:identifier>

in XML (or whatever). I don't know why DOI haven't tried to register
'doi' as a valid URI scheme. Clearly, having it registered would be
preferable.

My gut feeling goes against DCMI becoming a registry of identifier
schemes - I think we could do without the maintenance overhead. People
already have two places to register their schemes - as a
URI scheme and/or as a URN NID. Also, there may be some syntactic
advantages (particularly in RDF) in having *all* identifiers as URIs.

On the other hand, I guess that many 'users' of DC metadata will find

much more intuitive than

I presume we could use our 'fast track' process for registering these
things? Registering a URI scheme or a URN NID is more formal and long
winded - but this may not be a bad thing as it will encourage people to
think about what they are doing. The URI and URN registration processes
are outside of our control - so there may be an issue if a scheme that DC
people want to use is refused registration for some reason.

Note that, I assume, all valid schemes for Identifier are also valid for
Relation and Source. During the original discussion of qualifiers I'm
pretty certain that I proposed an initial list of schemes that would have
included

URI
DOI
ISBN
ISSN
SICI

Priscilla also proposed

Handle
URL

Ann Apps and I have since proposed

OpenURL

in http://www.ariadne.ac.uk/issue27/metadata/

As I say, my gut feeling is to stick with URI and force people to register
schemes externally to DCMI, either as new URI schemes or as URN NIDs or
both. But its not a very strongly held view and I wouldn't object to
registering stuff within DCMI.

Andy.

> Tom
>
> [1] http://www.ietf.org/rfc/rfc2288.txt
> [2] http://www.uri.net/urn-nid-status.html
>
>
> ---------- Forwarded message ----------
> Date: Mon, 20 Aug 2001 12:08:54 GMT1BST
> From: Ann Apps <ann.apps@MAN.AC.UK>
> To: DC-GENERAL@JISCMAIL.AC.UK
> Subject: Encoding schemes for resource identifier
>
> Dear All,
>
> Currently the only endorsed encoding scheme for dc:identifier is
> 'URI', and also for dc:relation and dc:source. It seems to me that
> you should be able to use other global standard identifiers as
> encoding schemes for these elements.
>
> Within the bibliographic citation area it seems natural to use
> identifiers like ISSN, DOI, SICI, OpenURL (possibly DOI is a URI?).
> In fact, people are already using these even though they are not
> endorsed by DCMI.
>
> I am sending this to dc-general, because I'm not sure which
> working group would consider it. It may fall into the domain of dc-
> citation. But I suspect there are other global identifiers in other
> domains.
>
> Interestingly, DOI and ISBN are given as examples of identifiers in
> the definition of dc:identifier in the DCMES, but are not given as
> qualifer encoding schemes.
>
> Should there be a list of recommended global standard encoding
> schemes, or should there be a more generic endorsement of any
> global standard? Having a fixed list means it will need extension
> when other standards are developed.
>
> Best wishes,
> Ann
>
> --------------------------------------------------------------------------
> 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: ann.apps@man.ac.uk WWW: http://epub.mimas.ac.uk/ann.html
> --------------------------------------------------------------------------
>

Andy
--
Distributed Systems and Services
UKOLN, University of Bath, Bath, BA2 7AY, UK a.powell@ukoln.ac.uk
http://www.ukoln.ac.uk/ukoln/staff/a.powell Voice: +44 1225 323933
Resource Discovery Network http://www.rdn.ac.uk/ Fax: +44 1225 826838