Re: Test methodology (was Re: Test #64: @rel in <head> using reserved XHTML value and safe, non-prefixed CURIE syntax)

Ivan Herman wrote:
> This actually raises another issue. Our current tests (including Manu's
> crazyivan) test whether a specific set of triples are in the output.
> However, none of the tests test whether the processors generates any
> unwanted extra triples. How would we do that?

If there were a NOT operator in SPARQL, or something to check
non-existence, that would be great. I couldn't find one in the 30
minutes I went looking for a solution.

Test #23 checks for non-existence in a very clever way by constructing
the XHTML in such a way that if an incorrect triple were to be
generated, the test would fail:

http://rdfa.digitalbazaar.com/rdfa-test-harness/

I don't know if this approach would work for every non-existence test,
but I doubt it.

Also, didn't we agree in one of the conversations that generating extra
triples is fine? So, should we really be writing tests that prevent
people from generating extra triples?

-- manu

-- 
Manu Sporny
President/CEO - Digital Bazaar, Inc.
blog: Over One Million Songs Available on Bitmunk
http://blog.digitalbazaar.com/2007/10/29/one-million-songs-on-bitmunk/

Received on Wednesday, 12 December 2007 16:40:47 UTC