Re: test schema issues

Eric Prud'hommeaux wrote:
> per
> 
> ACTION: EricP to fix test schema to match manifest with negative tests
> 
> I've been looking at the schema for tests. AndyS specifically kept
> http://www.w3.org/2001/sw/DataAccess/tests/test-manifest.n3 agnostic
> to the test harness:
>   mf:action rdf:type rdf:Property ;
>       rdfs:comment "Action to perform" ;
>       rdfs:domain   :ManifestEntry ;
>       # rdfs:range   ?? ;
>       . 

This is intentional - the manifest vocabulary should be usable for more than 
just query tests.  It could be some neutral class like mf:Test.

See also rdfs:seeAlso.

> 
> http://www.w3.org/2001/sw/DataAccess/tests/test-query.n3 also says
> nothing about the range of action. By inspection, the range appears to
> be something which packages a query and some data
>          mf:action
>             [ qt:query  <expr-1.rq> ;
>               qt:data   <data-1.ttl> ] ;
> or, in the case of the recent syntax tests, a query:
>          rdf:type   mfx:TestSyntax ;
>          mf:action  <syn-02.rq> ; 
>          ...
>          rdf:type   mfx:TestBadSyntax ;
>          mf:action  <syn-bad-01.rq> ; 
> 
> This strikes me as chaotic, and potentially non-mon,

example of non-monotonicity?

> depending on what
> assumptions people make about the range of action.

The manifest vocabulary is not making any assumptions - that's the point!

> 
> In the interest of rigorous modeling, I propose this hierarchy:
>                                   QueryTest
>                      /--------------/   \--------\
>           SyntaxQueryTest                   ExecutedQueryTest
>             /          \
> PositiveSyntaxTest  NegativeSyntaxTest
> 
>   qt:action rdf:type rdf:Property ;
>       rdfs:comment       "Action to perform" ;
>       rdfs:subPropertyOf mf:action ;
>       rdfs:range         qt:QueryTest;
>       . 
> 
> and update
> 
> to change
>          rdf:type   mfx:TestSyntax ;
>          mf:action  <syn-05.rq> ; 
> to
>          rdf:type   mfx:PositiveSyntaxTest ;
>          mf:action  [mf:query <syn-05.rq>] ; 
> 
> I'm trying to not change too many bytes here. Let me know if you think
> I should be more aggressiv.e

I don't see the point of chnaging class names just for the sake of changing 
names.  They were TestQuery/TestSyntax/TestBadSyntax/TestSerialization and I 
have code that uses that knowledge.

I had hoped that merging the additional properties would be a small matter of 
rerunning the Java generator (Jena's schemagen) and doing some trivial fixup. 
  Not having to check though the test harness because of redesign.

 Andy

> --
> -eric
> 
> office: +81.466.49.1170 W3C, Keio Research Institute at SFC,
>                         Shonan Fujisawa Campus, Keio University,
>                         5322 Endo, Fujisawa, Kanagawa 252-8520
>                         JAPAN
>         +1.617.258.5741 NE43-344, MIT, Cambridge, MA 02144 USA
> cell:   +81.90.6533.3882
> 
> (eric@w3.org)
> Feel free to forward this message to any list for any purpose other than
> email address distribution.

Received on Wednesday, 12 October 2005 09:53:54 UTC