Re: rr:constant with literal or datatype

* Richard Cyganiak <richard@cyganiak.de> [2014-11-16 15:16-0500]
> Eric, what would happen if Tomasz simply added the tests and put his version of the test suite up on GitHub? It would be convenient, as people could submit pull requests against it, with additional tests. But I'm worried about W3C copyright.

We have a cadre of lawyers commanding a secret police force, feared
and respected throughout the free world. Don't screw with our copyright.

I don't know. I'll dig into this with Ian, but my intuition is that
the copyright is just a way of enforcing the social contract of "we
asked members to a group; that group ratified some standard and some
tests to measure conformance (loosely) with that standard."

Our copyright let's folks redistribute without modification so I bet
it's a question of how you phrase the additional tests; if you do the
obvious thing of adding them directly to a copy of the manifest, you
insinuate that the WG has approved them; if you have a separate set,
they can be clearly marked as "not ratified but probably a good idea".
As frustrating as it may seem, it's probably a good idea to have a
clear line around the tests that have gone through a CR process.


> Richard
> 
> 
> 
> > On 16 Nov 2014, at 13:45, Eric Prud'hommeaux <eric@w3.org> wrote:
> > 
> > * tomasz.pluskiewicz@gmail.com <tomasz.pluskiewicz@gmail.com> [2014-11-16 14:31+0000]
> >> Thank you Richard
> >> 
> >> I have come to that conculsion after reading the specs carefully indeed.
> >> 
> >> Do you think such tests could be added to the suite? There a number of cases I can think of:
> >> 
> >> 1. Properly use the constant value as-is with dataype or language
> >> 2. Fail if a constant-valued has rr:datatype
> >> 3. Fail if a constant-valued has rr:language
> > 
> > Hi Tom, it sounds like you'd be the best one to write them. We can
> > either add them to the errata or figure out some sort of way to
> > communicate that there are other tests that weren't approved by the
> > WG. (In principle, I could just add them to the test suite, but that
> > seems to bypass an important social process.)
> > 
> > 
> >> Thanks,
> >> Tom
> >> 
> >> November 13 2014 7:10 PM, "Richard Cyganiak" <richard@cyganiak.de> wrote: 
> >>> Tomasz,
> >>> 
> >>>> On 7 Nov 2014, at 17:57, tomasz.pluskiewicz@gmail.com wrote:
> >>>> 
> >>>> Hi
> >>>> 
> >>>> I haven't found such example and I'm not sure what should happen if the R2RML mapping contained a
> >>>> constant with language tag or datatype. For example
> >>>> 
> >>>> [ rr:constant "someObject"^^<http://example.org/some#datatype> ]
> >>> 
> >>> There’s indeed no test case for this. So you’ll have to check the text of the spec.
> >>> 
> >>>> Should the datatype be ignored or used as if it was specified using rr:datatype?
> >>> 
> >>> As per R2RML §11.2:
> >>> 
> >>> [[
> >>> The generated RDF term of a term map for a given logical table row is determined as follows:
> >>> • If the term map is a constant-valued term map, then the generated RDF term is the term map's
> >>> constant value.
> >>> ]]
> >>> http://www.w3.org/TR/r2rml/#generated-rdf-term
> >>> 
> >>> And:
> >>> 
> >>> [[
> >>> The constant value of a constant-valued term map is the RDF term that is the value of its
> >>> rr:constant property.
> >>> ]]
> >>> http://www.w3.org/TR/r2rml/#dfn-constant-value
> >>> 
> >>> So, yes, the value of rr:constant is used exactly as is, including datatype if present.
> >>> 
> >>>> And more interestingly what should happen for the mapping below? Answering the first question
> >>> will
> >>>> suggest the answer.
> >>>> 
> >>>> [
> >>>> rr:constant "someObject"^^<http://example.org/some#datatype> ;
> >>>> rr:datatype <http://example.org/some#datatype>
> >>>> ]
> >>> 
> >>> [[
> >>> A term map that is not a datatypeable term map MUST NOT have an rr:datatype property.
> >>> ]]
> >>> http://www.w3.org/TR/r2rml/#dfn-specified-datatype
> >>> 
> >>> Following the definitions, you will find that constant-valued term maps (those with rr:constant) do
> >>> not have a term type, and therefore can’t have a term type of rr:Literal, and therefore are not
> >>> datatypeable.
> >>> 
> >>> Best,
> >>> Richard
> >>> 
> >>>> It looks like such cases aren't included in the test suite [1].
> >>>> 
> >>>> Kind regards,
> >>>> Tom
> >>>> 
> >>>> [1] http://www.w3.org/2001/sw/rdb2rdf/test-cases
> > 
> > -- 
> > -ericP
> > 
> > office: +1.617.599.3509
> > mobile: +33.6.80.80.35.59
> > 
> > (eric@w3.org)
> > Feel free to forward this message to any list for any purpose other than
> > email address distribution.
> > 
> > There are subtle nuances encoded in font variation and clever layout
> > which can only be seen by printing this message on high-clay paper.

-- 
-ericP

office: +1.617.599.3509
mobile: +33.6.80.80.35.59

(eric@w3.org)
Feel free to forward this message to any list for any purpose other than
email address distribution.

There are subtle nuances encoded in font variation and clever layout
which can only be seen by printing this message on high-clay paper.

Received on Sunday, 16 November 2014 20:38:14 UTC