UddpTestCases

From W3C Wiki

Sketches for some test cases for URI Documentation Discovery.

Assume AWWW treatment of information resources:

  • There is an IR that "consists of words and punctuation symbols [etc]"
  • Cars and dogs are not IRs

The most constrained case: hash URI, not an information resource, application/rdf+xml

 U = http://example.org/eq018#_
 where GET http://example.org/eq018 -> 200 Z
   and Z Content-type: application/rdf+xml
   and Z says that <U> is not an information resource
 Result: Z is a NUDC for U
 (uncontroversial)

What if a hash URI is said to name an IR, is that OK?

 U = http://example.org/eq018#_
 where GET http://example.org/eq018 -> 200 Z
   and Z says that <U> is an information resource
 Result: Z is a NUDC for U
 (not sure if anyone disputes this?)

Similarly for 303 NIR

 U = http://example.org/eq018
 where GET http://example.org/eq018 -> 303 Location: http://example.org/about/eq018
   and GET http://example.org/about/eq018 -> 200 Z
   and Z says that <U> is not an information resource
 Result: Z is a NUDC for U
 (uncontroversial)

How about 303 IR

 U = http://pubs.acs.org/doi/abs/10.1021/ed100697w
 where GET http://example.org/eq018 -> 303 Location: L
   and GET L -> 200 Z
   and Z says that <U> is an information resource
 Result: Z is a NUDC for U
 Result (for those who think 303 means NIR): Z is a NUDC for U, but is wrong

How about a 200... should break this case down

 U = http://example.org/eq018
 where GET U -> 200 Content-type: text/plain Z
 Result (status quo): 
   1. Z is (said to be) a representation of <U>
   2. ZZ is a NUDC for U
        where ZZ says that <U> is an information resource
 Result (Harry Halpin)
   1. Z is (said to be) a representation of <U>
   2. Z is a NUDC for U

How about 200 with RDF/XML...

 U = http://example.org/eq018
 where GET U -> 200 Z
 where Z has Content-type: application/rdf+xml
 Result (Ian Davis): Z is a NUDC for U

How about hash for non-http: URI?

 U = data:text/turtle,Z#_
 Result (JAR): text/turtle,Z is a NUDC for U
 Result (for those who think data: is out of scope): no agreement