Re: Some problems in test suite RDFa/manifests

Hi Richard

After a mercurial issue and rdfa parser selection … see my comments inline

On Feb 14, 2012, at 7:18 PM, Richard Cyganiak wrote:

> Boris,
> 
> Great work on the test suite – it's impressive to see what you've built up there in parallel to the other WG activities, and it's about time that your work gets a bit more attention from the rest of the WG!
> 
> So, I looked at the RDFa triples that I get from parsing the test cases document [1] with any23. I saw a couple of problems:
> 
> 1. There seem to be no rdf:type triples that mark the entities as databases, DM test cases, or R2RML test cases
Fixed.
> 
> 2. There seem to be no triples that connect test cases to the databases used by the test case
> 
Fixed.
> 3. There seem to be no links to the actual files (SQL, R2RML, nt, nq)
> 
Fixed.
> 4. The reviewStatus values are marked up as literals; they should be URIs:
> 
Fixed.
> 5. Sometimes, the specificationReference values are URIs (good), sometimes they are literals (bad), e.g., "#rules_table_triples"
> 
Fixed.
> In the original Turtle manifest files in the HG repository [2], some of these problems are present as well, but some are not, so there might be a problem with the RDFa template that you use when generating the HTML spec. (There might be problems with any23 too – always possible.)

I would prefer to use Ivan's suggestion:
http://www.w3.org/2007/08/pyRdfa/extract?uri=http%3A//www.w3.org/2001/sw/rdb2rdf/test-cases/
> 
> Another problem in the original Turtle files is that filenames like "directGraph.nt" and "create.sql" should really be relative URIs instead of literals so that everything is properly hyperlinked.

I'll include this in my TO DO.

Thanks again Richard

Boris


> 
> Best,
> Richard
> 
> 
> [1] http://www.w3.org/2001/sw/rdb2rdf/test-cases/
> [2] https://dvcs.w3.org/hg/rdb2rdf-tests/file/default