Issue/Derivative resources

Alan had written:

TBD:

1. Explanation that format records simple URI are "authoritative" - explain sense of authoritative: We have http://sn.org/uniprot/P04637, http://sn.org/uniprot/P04637.rdf, http://sn.org/uniprot/P04637.xml. Where do these encodings (the RDF and XML) come from. Proposal: They come from the publishers of uniprot ("the authority"). If there are other "non authoritative" encodings, e.g. bio2rdf RDF encoded uniprot, then they would have a different URI, e.g. http://sn.org/bio2rdf/uniprot/P04637.rdf
2. URI form for alternatively sourced (non - authoritative) encodings

JAR writes:

For any given Shared Names URI, the project decides what resource is named by the URI. At any given time Shared Names may recognize one or more entities as being "authorities" for a resource. An "authority" is consulted when needed on questions relating to that resource. "Consulting" might mean automatic means (redirects) or manual means (a curator checking an authority's web site to find out where something is).

Each idspace has one or more designated authorities consulted for URIs in that idspace. The idspace's authorities are the authorities for any given accession (= record) in some idspace. For example, Uniprot might be an authority for the 'uniprot' idspace, and would be consulted regarding http://sharedname.org/uniprot/P04637. For a given idspace and accession number there are auxiliary resources for which the idspace's authority is authoritative, e.g. for the Uniprot XML page http://sharedname.org/uniprot/P04637.xml we consult Uniprot, as well as derivative resources that have their own authorities, such as RDF generated by Bio2RDF or Neurocommons.

The about-record for http://sharedname.org/uniprot/P04637 can link to as many derivative resources as we like. Given that, the derivative resources may also have or not have Shared Names URIs. If they do we might have http://sharedname.org/uniprot/P04637.bio2rdf-rdfxml , or some other syntax.

What should those URIs look like? Should there be some documented way of marking that these resources are derivative and not primary (i.e. coming from the authority)? Should they go in their own idspaces?

N.b. "author", "authority", "location", and "source" are all different things. The author would be who, in fact, is responsible for creating the resource that Shared Names means to name by a given URI. An authority is an entity that Shared Names trusts to answer questions about the resource in question. And the location is just the place where any content is stored; it may bear no relation to either the author or the authority. (Of course when we communicate on the web with the location, we endow it with authority, since we will generally believe what it tells us.) "Source" is ambiguous as it could mean either ultimate source (author) or proximate source (location).

Apologies for using the word "resource" - the sense is that of RFC 2616, and I couldn't come up with a better word.