Re: Roman Numeral test Was: Bug: "A value disjunction that encounters a type error on only one branch will return the result of evaluating the other branch."

On Wed, Aug 10, 2005 at 09:24:54AM -0400, Tim Berners-Lee wrote:
> 
> Here I understand Yosi coded up cwm's SPARQL implementation with  
> special SPARQL-compatible builtins written so as to match the spec.   
> That doesn't mean that the spec is right or sensible or logical.  :-)

I have spoken with TimBL and I believe that changing the text 

[[
A value disjunction that encounters a type error on only one branch
will return the result of evaluating the other branch.
]]

to

[[
A value disjunction that encounters a type error on only one branch
will return TRUE if the other branch is TRUE and an error if the other
branch is FALSE.
]]

will satisfy him.

The truth table below describes the changes:

left	  right	result
T	  T	T
T	  F	T
F	  T	T
F	  F	F
T	  E	T
F	  E	F/E	Was FALSE, now Error
E	  T	T
E	  F	F/E	Was FALSE, now Error
E	  E	E

I find either easy to implement. Tim's motivation is to give the rule

  F || X ==> X
  False OR X evaluates to X

priority. Opinions?

http://www.w3.org/2001/sw/DataAccess/rq23/#func-logical-or


> On Aug 8, 2005, at 15:03, Dan Connolly wrote:
> 
> >On Mon, 2005-08-08 at 13:42 -0500, Dan Connolly wrote:
> >
> >>FYI, there's another test case available to study:
> >>
> >>Roman numeral test Dave Beckett (Monday, 8 August)
> >>http://lists.w3.org/Archives/Public/public-rdf-dawg/2005JulSep/ 
> >>0228.html
> >>
> >>->
> >>
> >>http://www.w3.org/2001/sw/DataAccess/tests/data/ValueTesting/roman.rq
> >>http://www.w3.org/2001/sw/DataAccess/tests/data/ValueTesting/roman.n3
> >>http://www.w3.org/2001/sw/DataAccess/tests/data/ValueTesting/roman- 
> >>result.n3
> >>
> >>
> >>Yosi, if you could look at that soon, I'd appreciate it.
> >>
> >
> >I just looked at it with that cwm sparql server on mr-burns, yosi.
> >
> >It gives 0 results, which agrees with the last call design
> >(and disagrees with the roman-results.n3 sketch).
> >
> >TimBL, can you confirm that cwm is giving 0 results by design?
> >
> >
> >
> >>The valueTesting issue (http://www.w3.org/2001/sw/DataAccess/ 
> >>issues#valueTesting )
> >>is on the agenda for tomorrow's teleconference, and it would
> >>be nice to have input from the cwm/swap project in hand.
> >>
> >[...]
> >
> 

-- 
-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.