Re: How to specify prefixes for test case conclusions

Chris Welty wrote:
> This fulfills ACTION-633.
> In test cases, we have the (very minor) problem of how to advise
> implementors to treat the conclusion specifications - are they different
> documents or the same document as the test case premise.  If they are

They are not documents.  They are simply condition formulas.

> the same document, then it is possible to reuse the namespace prefixes
> from the premise and not re-declare them in the conclusion.  Also, if

We can have this as a convention in the test case specification:
CURIES in the conclusion are understood as full IRIs, expanded according
to the prefix definitions in the premise.
Another possibility would be to have a number of prefix definitions on
the test case level, which would be used to expand CURIES both in the
premise and conclusion.

> they are the same document then the "Local_Constant" test case [1] will
> not work properly.

I cannot see how this would work.  The BLD syntax does not have a place
for conclusions in documents.

Best, Jos

> </chair>
> I suggest we approve the convention that the conclusion is the same
> document, and rewrite the local constant test case to use imports to
> achieve its effect. This convention appears to have been adopted in most
> test cases anyway.
> <chair>
> I am extremely wary of any proposal to change BLD in order to address
> what seems like a fairly simple problem.
> -Chris
> [1]

Jos de Bruijn  
No one who cannot rejoice in the discovery of
his own mistakes deserves to be called a
  - Donald Foster

Received on Friday, 30 January 2009 09:16:48 UTC