> From [log in to unmask] Mon Feb  4 16:59 MET 2002
> MIME-Version: 1.0
> Content-transfer-encoding: 7BIT
> X-Scanner: exiscan *16XlW4-000MnZ-00*Nm/Qm.w3FeA* (Manchester Computing,
>            University of Manchester)
> Date:         Mon, 4 Feb 2002 15:58:49 GMT1BST
> From: Ann Apps <[log in to unmask]>
> Subject:      Re: Guidelines for encoding identifiers in              Dublin
>               Core metadata
> Comments: To: [log in to unmask]
> To: [log in to unmask]
>
> Andy,
>
> Some comments on your Guidelines for encoding identifiers in DC
> document:
>
> I guess that part of your reasons for introducing these guidelines is
> to remove the need to add/register new schemes for dc:identifier,
> and simliar elements. If they can all be encoded as URIs, that
> won't be necessary.

The question is not, whether one could recode identifiers to URI's -
the issue is, whether a general (re)-code to URI is GIVEN for the
specific scheme.


> However, I'm a bit concerned about whether
> the DC community at large will actually follow these guidelines for
> common identifiers. People are already writing things like:
> <dc:identifier scheme="ISSN">1234-5678</dc:identifier> and
> <dc:identifier scheme="DOI">10.1060/abcdefgh</dc:identifier>
> I'm not sure that they will change to:
> <dc:identifier scheme="URI">urn:issn:1234-5678</dc:identifier> and
> <dc:identifier scheme="URI">doi:10.1060/abcdefgh</dc:identifier>
> even if it is more accurate. Maybe this doesn't matter!

It does! You immediately see the difference in dumb down:
10.1060/avcdefgh ??  versus
doi:10.1060/abcdegh.

Both constructions can be harmful with XML: You could have any UTF-8
coded character in a doi. In particular those, which are reserved in XML
for genuin XML processing.



>
> Following on from that, I can envisage situations within an
> application where I would need something like:
> <dc:identifier scheme="ISSN">1234-5678</dc:identifier>
> for instance, if I needed to index all ISSNs. But I guess in this case
> I could use an application-specific element.

What you do privately is not the concern. What you send to somebody else
matters. What you mean with an application specific element?

my:issn ??




>
> These are guidelines for encoding 'commonly used identifiers'.
> Maybe the word 'standard' should be used somewhere. All the
> ones in the document are, I think, standards. But then I suppose
> there will be some common ones which aren't.
>
> Is this document meant to be examples of how to encode such
> identifiers? Or is is meant to be a definitive list?
>
> If it is guidelines/examples then something should be said about
> other identifiers not mentioned in the document. The problem is
> that it's not easy to see what the criterion is for encoding ids this
> way. If it is the list maintained by IANA then that doesn't seem to
> work too well - some of these are not registered (and the document
> is dated 2001-08-20 which doesn't give me too much faith in their
> registration process!).

IANA is authoritive for registered schemes.



> What about ids which are not urns - I believe
> BICI is one?

What is BICI??

>
> If this is a definitive list, then there will be problems with
> extensibility, registration, maintenance, etc. I immediately wonder
> how to use BICI, OAI, PubMed, etc. Would ids such as oclc-
> number be in scope?

No this can't be a definitive list. Each library system is proud of
it's own identifier scheme.


> You certainly wouldn't want to update the
> document every time some other significant identifier appeared. So
> you'd have to think about some sort of registration - maybe you
> envisage the IANA list as providing this but I suspect we'd want ids
> not on that list.
>
> Should DC be looking at IRI's rather than URIs with Unicode
> inclusion? http://www.w3.org/International/O-URL-and-ident.html
>
> I think you may need to include something about escape encoding
> in this document. SICIs certainly need it - and your example is
> escape-encoded.
>
> Being really picky :), I note that you HTML examples use the form
> 'DC.Identifier' rather than the preferred 'dc:identifier'!
>
> Does the 'handle system' exist for anything other than DOIs? I don't
> think it's in very common usage. I'm not sure it's too helpful to
> include it.

DOI's are those handles, which start with a 10. OR?

rs

>
> I hope some of this is useful. Your document looks like a good
> start for guidelines in this area which are very obviously needed.
>
> 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: [log in to unmask]  WWW: http://epub.mimas.ac.uk/ann.html
> --------------------------------------------------------------------------
>