Am 02.02.2015 um 13:18 schrieb Martin de la Iglesia:
> I just found this old thread because I was looking for a solution to a
> similar problem: I would like to indicate both the type of an identifier
> in <idno>, as well as a preferred resolver for the identifier. I was
> thinking of something like
> <idno type="GND" xml:base="http://d-nb.info/gnd/">118575449</idno>
> but I'm not sure whether that conforms to the definition of @xml:base
> ("provides a base URI reference with which applications can resolve
> relative URI references into absolute URI references"). The idea of
> @xml:base seems to be that it applies to attribute values of the
> data.pointer type only. What do you all think?
I think the general approach makes sense, but I share your concerns with
regard to the semantics of xml:base.
> On a related note, does anyone else feel that the Guidelines example
> <idno type="DOI">http://dx.doi.org/10.1000/123</idno>
> is misleading? After all, "http://dx.doi.org/" is the resolver URL, not
> part of the DOI proper.
While there are cases where the URL might be the identifier, and thus
you could just use it as content of <idno>, this is probably not true
for both DOI and GND. They are identifiers that you can look up using a
URL, but that don’t contain the lookup scheme themselves.
This seems to be similar in principle to the way cRefs work, except that
you don’t use the @cRef attribute. But the situation is the same: You
have a canonical referencing system that allows you to translate a
reference into a URL.
It might be a bit overkill, but I could imagine a listIdnoDef/idnoDef
pair in addition to listPrefixDef/prefixDef and refsDecl/cRefPattern:
<idnoDef type="GND" matchPattern="(.*)"
Or one could extend the cRefPattern mechanism by allowing to specify
additional elements/attributes that are treated as canonical references
in addition to @cRef, something like:
<cRefPattern match="idno[@type='GND']" matchPattern="(.*)"
(This example uses an XSLT-like @match that might not be perfectly in
line with the semantics of TEI @match, it is just meant to illustrate
the general idea.)
I don’t know if you use case would justify such an addition, so these
are just my 2¢.
Dr. Frederik Elwert
Centre for Religious Studies
Phone +49(0)234 32-24794