On Wed, 6 Mar 2019 at 15:29, Denenberg, Ray <[log in to unmask]> wrote:

Can I bring the discussion back to the original question, whether a URI, supplied as the value (rdf:value) of bf:identifiedBy, should be encoded as a literal or an actionable URI. 


I believe the question  is probably irrelevant, because, I still believe, there is no practical purpose served by doing so.


Let me ask this:  if it is to be an actionable URI,  what do you expect to get upon dereferencing it?

·         Do you expect to get a description of the resource?

Hi Ray:

By dereferencing a URI, I hope to get:
  * useful information, using the standards (RDF*, SPARQL)
  * links to other URIs, so that I can discover more things

Let's go with the ISSN example URI for the "Canadian Journal of African Studies / Revue canadienne des études africaines" (http://issn.org/resource/ISSN/1923-3051), even though issn.org may be a bit confusing because it adopts some BIBFRAME for its vocab along with schema.org and DC.

Using the ISSN URI (or other URIs, for other types of identifiers) would allow me to rely on an external source of authoritative metadata for different identifiers.

The basic ISSN identifier literal "1923-3051" is readily available from that URI via its schema:issn property.

The information includes other properties and relationships, such as ISSN-L. These may help me with cataloguing efforts. They may also serve search and display operations.

As a bonus, using the ISSN URI also lets me, or anyone else, publish statements about that resource. The linked data produced by issn.org does not include the title in French, which my multilingual (French and English) library really wants. The linked data produced from the LCCN (79018614) includes the French title as one of the variant titles, but fails to mark it as French, which makes it hard to work with for search & display purposes in an automated system. So if I do the work to untangle those two titles and state them with proper language tags, I could publish:


  bf:mainTitle "Canadian Journal of African Studies"@en;
  bf:mainTitle "Revue canadienne des études africaines"@fr;

(Yes I know BF wants <URI> bf:title [ bf:mainTitle "title"@lang" ] but I'm following ISSN's pattern here)

Anyone who trusts my library can benefit from those statements at the level of the ISSN URI. And if I have print as well as electronic versions of the resource, I can follow the ISSN-L relationship from the information published by the ISSN URI and populate those title statements accordingly

Otherwise, in the approach that is used by current BIBFRAME, I would have to do something like:

  bf:identifiedBy [
    a bf:Issn ;
    rdfs:value "1923-3051" ;
  ] ;

    bf:mainTitle "Canadian Journal of African Studies"@en;
    bf:mainTitle "Revue canadienne des études africaines"@fr;

So in that scenario, anyone consuming the data that I publish is going to have to do additional work to construct a URI for the ISSN. I'm concerned that this latter approach does not make our data much more useful outside of the library realm than looking for 022 $a in a MARC record.

  We’ve established that that’s not an appropriate use of an identifier.

It has been asserted, but I'm not convinced that point has been established. In your message on February 22nd, you said:

'Issn: 1084-5305' is a piece of information included in a Work description, like subject, genre, etc. That’s a big difference, a description of a resource vs. a piece of information in the description.  And I would say that bf:identifiedBy is intended for the latter.

This seems to suggest that "a piece of information" (ISSNs, subjects, genres) should be a literal instead of a URI. But bf:subject and bf:genreForm do take URIs directly, so that assertion confuses me.

After all of this, it's still not clear to me what the intention of bf:identifiedBy is, and why that intention prescribes a literal value.

Other than the BIBFRAME RDF Schema + vocabulary list, is there a design document that lays out the rationale for these and other BIBFRAME design decisions, and how they are envisioned to be used? I feel like I've had to piece together the little that I know from reading the mailing list, closely analyzing the XSLT conversions, and gleaning what I can from articles, presentations, and Slack discussions.