- From: Steve Harris <S.W.Harris@ecs.soton.ac.uk>
- Date: Fri, 11 Nov 2005 16:25:34 +0000
- To: DAWG public list <public-rdf-dawg@w3.org>
Whilst tidying up my EBV implementation yesterday, I noticed an
inconsistency between the tests and the wording in rq23.
rq23 (11.2.2 Effective Boolean Value) has that all operands are true,
except:
unbound.
an xsd:boolean with a FALSE value.
a 0-length untyped RDF literal or xsd:string.
any numeric type with a value of 0.
an xsd:double or xsd:float with a value of NaN
However,
http://www.w3.org/2001/sw/DataAccess/tests/#test-boolean-effective-value-optional
implies that FILTER("foo"^^:unknown) fails, either through a type error or
just because its false, but the spec doesn't back this up.
On IRC Andy pointed out that assuming all things of unknown types are true
is a bit closed-wordly, I'm not aure if I agree with that or not, but it
needs more clarity.
- Steve
Received on Friday, 11 November 2005 16:25:51 UTC