W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > March 2003

Re: Test cases for literal equality?

From: Dave Beckett <dave.beckett@bristol.ac.uk>
Date: Thu, 13 Mar 2003 11:32:07 +0000
To: Graham Klyne <gk@ninebynine.org>
cc: w3c-rdfcore-wg@w3.org
Message-ID: <2198.1047555127@hoth.ilrt.bris.ac.uk>

>>>Graham Klyne said:
> 
> Do we have any test cases for dealing with literal equality?
> 
> In particular, I'm wondering if the recent discussion of XML literals and 
> canonicalization will have any effect of the interpretation of language 
> tags for typed literals.  Currently, if I have the details right, typed 
> literals with different language tags are distinct values in the abstract 
> graph, but always denote the same thing, with the exception of XML 
> literals.  Plain literals are language-tag sensitive.  What about xsd:string?

xsd:string is a datatype in the XSD specification and from what I
recall, RDF doesn't use it - no RDF literal is an xsd:string nor has
one as a part, although the lexical form definition is compatible
with it.  A quick grep in the concepts WD confirms this as far
as I can tell.  So we don't need to test xsd:string comparisons.


> I didn't find any entailment tests covering this area.

The nearest we have is
  http://www.w3.org/2000/10/rdf-tests/rdfcore/rdf-charmod-uris/

which checks that the literals from the syntax (RDF/XML) correctly
get into the RDF graph (as N-Triples)

Slightly related, there is the N-Triples test, which exercises the
N-Triples syntax by having many forms.  It should entail itself.
Linked from http://www.w3.org/TR/rdf-testcases/#ntrip_tests
to http://www.w3.org/2000/10/rdf-tests/rdfcore/ntriples/test.nt
and includes many XML Literal forms; which I guess I should check are
appropriately canonical.

> While I'm on this topic, I assume it's clear that the following describe 
> the same graph:
> 
> [[
> <rdf:RDF xmlns:rdf='http://www.w3.org/1999/02/22-rdf-syntax-ns#'
>           xmlns:ex='http://www.example.org/xxx/a>
>    <ex:subj>
>       <ex:prop>abc</ex:prop>
>    </ex:subj>
> </rdf:RDF>
> ]]
> 
> and
> 
> [[
> <rdf:RDF xmlns:rdf='http://www.w3.org/1999/02/22-rdf-syntax-ns#'
>           xmlns:ex='http://www.example.org/xxx/>
>    <ex:asubj>
>       <ex:aprop>abc</ex:aprop>
>    </ex:asubj>
> </rdf:RDF>
> ]]
> 
> ...

(With a bit of fixing up - missing 's)

Yes.

> On my attempt to use it for real, I'm finding the test cases document 
> difficult to follow, for two main reasons:
> 
> (a) it is formatted as a very wide document, such that the full width will 
> not fit on my display.
> 
> (b) the test cases are not obviously organized by functional groups, but by 
> issue.  While it made sense to use an issue-based organization for 
> assembling the tests, I think a more functionally oriented arrangement 
> would better serve the goal of clarifying questions about deign intent.

I'll leave this for Jan to say more on but I feel it would be a much
longer document without the 5 column tables to summarise the tests
and the test cases are for the issues - that is where the discussion
lies.  The test cases WD shouldn't try to duplicate that.

Dave
Received on Thursday, 13 March 2003 06:32:54 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:56:13 EDT