Re: Test Cases for Issue http://www.w3.org/2000/03/rdf-tracking/#rdf-ns-prefix-confusion

Dave Beckett wrote:
> 
> I have fixed a few bugs, added more tests and bad test cases for
> non RDF documents.

Very nice work.

A quibble, just to show I read it carefully:

This isn't an anonymous node; it's just a relative identifier:

  @prefix : <#> .
  :genid <http:...> <...>.

I'm proposing that we use
	_:genid <http:...> <...>.

i.e. no @prefix, just _:name. TimBL hacked support for it yesterday;
let's see if it works...

yup...

$ python2 cwm.py --quiet --rdf
http://ilrt.org/people/cmdjb/2001/05/rdf-ns-prefix-confusion/test01.rdf
--n3=spart --bySubject
    
    <http://example.org/resource1/>     <http://example.org/property>
"bar" .

hm... test2 doesn't work as well; cwm.py doesn't believe in bagID.

in test4.n3, you've got a relative reference in your expected results,
resulting from rdf:ID="foo". We're going to have to specify the
base URI as part of test inputs so that we can absolutize everything
consistently.

Aha... test5 allows me to show what I'm after:

$ python2 cwm.py --quiet --rdf
http://ilrt.org/people/cmdjb/2001/05/rdf-ns-prefix-confusion/test05.rdf
--n3=spart --bySubject
    
    <http://example.org/resource1/>     <http://example.org/property>
_:0 .
    
    _:0     <http://example.org/property2> "bar" .

I'm using

  http://www.w3.org/2000/10/swap/cwm.py
  $Id: cwm.py,v 1.52 2001/05/30 23:08:24 timbl Exp $

re test6: whoa! I didn't realize rdf:type as an attribute
was special syntax; i.e. that it takes a URIref rather
than a literal. None of the code I've written groks that.
Learn something new every day, I guess.


> These are testing the three parts of this issue:
> 
> 1. Tests 01-08 and bad tests 01-08
>   The following RDF attributes are required to have an rdf: prefix:
>     about aboutEach aboutEachPrefix
>     ID bagID type resource parseType
> 
> 2. Test 09 and bad test 09
>   Namespace qualification MUST be used for all property attributes.
> 
> 3. Tests 10-14
>   Non-prefixed RDF elements (NOT attributes) are allowed when a
>   default XML element namespace is defined with an
>   xmlns="http://www.w3.org/1999/02/22-rdf-syntax-ns#" attribute.
> 
> All the tests have an input RDF/XML file test<n>.rdf with the
> required output triples in RDF/n3 test<n>.n3 and the bad tests are in
> test<n>-bad.rdf
> 
> They have all been tested with SiRPAC and cwm/n3 as far as is
> possible (cwm doesn't seem to like aboutEach*).
> 
> I am attaching them cloaked to this message but they should move
> somewhere well known on w3.org where we can tweak them, add
> explanatory html, pointers to code etc. when we get those.
> 
> Dave


-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/

Received on Thursday, 31 May 2001 11:25:42 UTC