[UPS] more on identifiers & display

[herbert van de sompel]

Hi Carl & group,

I would like to add just one more thing to the whole discussion about
display URL etc..., which might be constructive when thinking about a
solution.  

- The first consideration is that we are discussing what a URL or ID
points at, basically is it a display page with metadata & links to
full-text or is it versions of the full-text, ...  

- The second consideration is that there definitely is an identifier for
the metadata itself, since Dienst will use that one to actually harvest
the metadata.  In most cases, I guess, the same identifier could also
lead to a metadata display page. I think this is an important identifier
too, which we don't seem to have explicitely in our meta data set.

- The third consideration is that there seems to be cases where there is
no full-text location (abstracts only stuff)

So, another solution could be to explicitely have 2 Location fields:

* Metadata location [not repeatable; mandatory since at least an ID must
exist here]

* Content location [repeatable; not mandatory]

and each of those could have link-to-URL and/or ID.  In some cases,
Metadata ID and Content ID will be the same, and it is the construction
of the link-to-URL that will make a distinction between showing metadata
or full-content.

I would like to add to this, that what I suggest here, corresponds to
the existing situation in the established communication mechanisms,
where we have:
- "accession numbers" for records in A&I databases (PubMed ID, for
instance)
- other "identifiers" (DOI, SICI-like constructs, SLinkS algorithmic
URLs) for the corresponding content.  

greetings

herbert