Proposed response to Sandro, manifest changes.

On Wed, 12 Nov 2003, Dan Connolly wrote:

> On Wed, 2003-11-12 at 04:14, Jan Grant wrote:
> [...]
> > So "where we're at" is that there are several options.
>
> Please pick one and propose it.
>
> i.e. draft a response to Sandro; then the chair
> will let you know if it needs WG input.

In order to address all of Sandro's comments on the closed-world
assumptions in the test case manifest, the following changes will be
made to the manifest format:


For parser tests, exactly one each of the test:inputDocument and
test:outputDocument elements must be present. The test case document
will be updated with words to ensure exactly this.

For entailment tests, the test:conclusionDocument is mandatory and must
be present exactly once for each test case.

Since entailment tests exist which take multiple premise documents, the
test:premiseDocument is renamed and becomes list-valued, and is
represented as follows:

<test:premiseDocuments rdf:parseType="Collection">
  <test:NT-Document rdf:about="..."/>
  <test:NT-Document rdf:about="..."/>
  ...
</test:premiseDocuments>

This form is used regardless of the number of premise documents (there
may be one or more).

The list of entailmentRules is replaced by a single-valued element which
must be present once and is mandatory. It takes the following form:

<test:entailmentRules
rdf:resource="http://www.w3.org/2000/10/rdf-tests/rdfcore/testSchema#Simple"/>

and can take values in the test: namespace:
	test:Simple	to indicate simple entailment
	test:RDF	to indicate rdf entailment
	test:RDFS	to indicate rdfs entailment
	test:Datatypes	to indicate rdfs with additional datatypes

No value is present for "RDF + datatypes(...)" because semantics builds
on RDFS to describe datatyped interpretations.

If the "Datatypes" entailment rules are chosen, then a list-valued
test:datatypeSupport declaration MUST be present. Its form is as
follows, with zero or more datatypes declared.

<test:datatypeSupport rdf:parseType="Collection">
  <rdf:Description rdf:resource="http://www.w3.org/2001/XMLSchema#integer" />
  ...
</test:datatypeSupport>

There is no declaration necessary for the rdf:XMLLiteral datatype since
it is required by RDFS.

Finally, the test:document element in the miscellaneous test cases
is replaced with a collection-valued test:documents.


The test case document will be amended with an updated description of
the format; the manifest descriptions of all test cases will similarly
be updated.


[OPTION, which MAY be added to the proposal if WG considers necessary,
is to change the URL of the large manifest file to something like
.../ManifestV2.rdf]





-- 
jan grant, ILRT, University of Bristol. http://www.ilrt.bris.ac.uk/
Tel +44(0)117 9287088 Fax +44 (0)117 9287112 http://ioctl.org/jan/
...You're visualising the _duck_ taped over my _mouth_..?

Received on Thursday, 13 November 2003 04:11:12 UTC