Re: Roman numeral test (issues#valueTesting)

On Mon, 2005-08-08 at 17:26 +0100, Dave Beckett wrote:
> I've added something to try to address my action from last week:
>   ACTION DaveB: make 'XXI'^^:romanNumeral  = 21 and points nearby into 
>   test cases (or ask questions in email).
> See:
> I'm not sure if it's quite what's wanted as you could just do
>  WHERE { anything you like.
>          FILTER ( 'XXI'^^rome:Numeral = 21 )
>  }
> but instead I put the roman numeral constant in the data and the 21 in
> the query.

Let's see...

SELECT ?legion
  ?legion roman:legionNumber ?n .
  FILTER ( ?n = 21 )

and the result says that yes, there's one match; the case of
?n binding to "XXI"^^roman:Numeral does meet the FILTER constraint.

Now as I understand the last call design (based on phone discussion
with EricP, not based on careful study of the text), it says there
are no matches.

So yes, this proposed test captures the issue I want discussed
quite nicely, Dave.

Gold star to anybody who comes up with suggested text for a design
that agrees with this test case so that we can compare it to the
last call text.

I also understand that the extendedType-eq-pass-result.n3 materials
we just approved follows the last call design in this regard.

EricP and I talked about various design options and I gather he's
doing some hacking around at least one of them.

Dan Connolly, W3C
D3C2 887B 0F92 6005 C541  0875 0F91 96DE 6E52 C29E

Received on Monday, 8 August 2005 18:58:22 UTC