I · 11

Identifiers

The info: URI scheme (RFC 4452) lets an OpenURL carry identifiers from common namespaces — DOI, PMID, ISBN, arXiv, source systems — in a uniform way.

Whenever an entity in a ContextObject is known by a stable identifier — a DOI, an ISBN, a PubMed ID — that identifier can be carried in the <entity>_id parameter directly, often eliminating the need for descriptive metadata. OpenURL expresses these identifiers as URIs in the info: scheme (RFC 4452).

The info: URI scheme

RFC 4452 defines a URI scheme of the form info:<namespace>/<identifier>. Namespaces are registered, and each names a community-managed identifier system. Within OpenURL, two kinds of info: URI matter:

  • info:ofi/<item> — items registered by the OpenURL Framework Initiative itself: formats, transports, encodings, profiles.
  • info:<ns>/<value> — identifiers from other registered namespaces (DOI, PMID, arXiv, and so on), used to identify entities.

Common identifier forms

IdentifierForm
DOIinfo:doi/10.1103/PhysRev.47.777
PubMed IDinfo:pmid/12345678
ISBN (info)info:isbn/9780486417134
ISBN (urn)urn:ISBN:9780486417134
ISSNurn:ISSN:0031-899X
OCLC numberinfo:oclcnum/1234567
arXiv IDinfo:arxiv/2103.00010
Source IDinfo:sid/example.com:database

Both the info: and urn: URI schemes are accepted by resolvers. The info: forms are the ones registered in the OpenURL Registry; the urn: forms predate the Registry and remain widely used.

Where identifiers go

Identifiers are placed in the <entity>_id parameter for the relevant entity. For the Referent, this is rft_id:

rft_id=info:doi/10.1103/PhysRev.47.777

For the Referrer — identifying the source system that emitted the ContextObject — the conventional pattern is rfr_id=info:sid/<source>:

rfr_id=info:sid/google.com:scholar

The info:sid/ namespace is reserved for identifying source systems; the value following the slash is by convention a reverse-DNS-style identifier for the source, with an optional colon-delimited sub-identifier.

Identifier and metadata together

An OpenURL may carry both an identifier and a full set of descriptive metadata. Doing so is recommended: resolvers can use the identifier when available and fall back to metadata when the identifier is unknown to their knowledge base. The two are not mutually exclusive.

rft_id=info:doi/10.1103/PhysRev.47.777
&rft.atitle=Can+Quantum-Mechanical+Description+of+Physical+Reality+Be+Considered+Complete%3F
&rft.jtitle=Physical+Review
&rft.aulast=Einstein
&rft.date=1935
&rft.volume=47
&rft.spage=777

Identifier-only requests

An OpenURL containing only an identifier and a format declaration is sufficient for a resolver that knows the identifier's namespace. Many resolvers will accept and resolve a request of this minimal form:

https://resolver.example.edu/openurl?
ctx_ver=Z39.88-2004
&rft_val_fmt=info:ofi/fmt:kev:mtx:journal
&rft_id=info:doi/10.1103/PhysRev.47.777

Sources