Theo,

> Thanks. For the time being we will use the tel namespace
> until the collection description profile is stable and we
> will use "hasService" instead of identifier.

Great. Based on where we have got to here, that sounds like the most
"compatible" approach! And hopefully we can firm up our proposal here
soon.

So... I'm pondering the suggestion that we use encoding schemes like
cld:SRUbaseURL and cld:Ztarget associated with the value of the
hasService(-or-whatever-we-agree-on) attribute.

I must admit that my first reaction to seeing this use of encoding
schemes was that this convention was stretching/overloading the encoding
scheme mechanism to carry information that is really an attribute of the
target resource (the Service). i.e. it's really "related metadata" in
the terms of

http://www.ukoln.ac.uk/metadata/dcmi/abstract-model/

I'm not at all sure I'm right on this, and I'm not sure I can really I
can back up that reaction, but I'll try....

(N.B. I'm _not_ arguing that the information is not useful to an
application using collection/service metadata but I'm trying to ensure
that we aren't abusing the DC metadata model in how we propose to
represent this information.)

The Usage Board definition of an encoding scheme is

> An Encoding Scheme provides contextual information or parsing
> rules that aid in the interpretation of a term value.
> Such contextual information may take the form of controlled
> vocabularies, formal notations, or parsing rules.  If an
> Encoding Scheme is not understood by a client or agent, the
> value may still be useful to a human reader.  There are two
> types of Encoding Scheme:
[snip]
> Syntax Encoding Schemes indicate that the value is a string
> formatted in accordance with a formal notation, such as "2000-01-01"
> as the standard expression of a date.

http://dublincore.org/usage/documents/principles/

In the case under consideration here, it seems to me that the "formal
notation/parsing rules" for interpreting the base URL of an SRW/SRU
target or an OAI repository are those of the URI and HTTP specs. AFAIK,
the SRW/SRU and OAI PMH specs don't specify any additional formal
notation/parsing rules for the interpretation of the base URL of the
service? So on that basis it might be argued that we shouldn't _need_ to
apply a separate encoding scheme to that value?

On the other hand, there might be an argument that in applying the
encoding scheme you are providing "contextual information... that aids
in the interpretation of the value"?

It may be useful to note that the Citation WG have a list of Identifier
Encoding Schemes here

http://epub.mimas.ac.uk/DC/citids.html

and that may be worth looking at.

(N.B. I think the reference to "OAI" in that document is to the OAI
identifier as defined in the OAI-PMH spec, not to the base URL of an OAI
repository)

Ann and Andy both have a much better grasp of identifier issues than me,
so I'd really value their opinions on this question.

But my gut feeling (at the moment!) is that using an encoding scheme
like cld:SRUbaseURL is using the encoding scheme mechanism to carry
information that should be expressed in some other way (as an attribute
of the Service).

Cheers

Pete