Re: Ideas for XSLT/Xpath Test Suite

["Carmelo Montanez" <carmelo@nist.gov>]

> Since there's many instances of alternative behavior defined in the
> XSLT Recommendation, just "YES" or "NO" doesn't give you enough
> information to know when to include the test and when not to include
> it.
>
> Perhaps an early activity should be to find where alternative
> behaviour is allowed and identify each of the alternatives?  We will
> then have names for some of the information that goes into the markup
> and/or data in the catalogue instances and rendition control
> instances.
>
> Regards,
>
> Tony Graham
>

I agree that we need to find which instances may produce different
behaviours and the sooner we identify those the better, however the
specifications file may not be
proper place for it.  Remember the toggling and switching is something that
the user does, not the test developer.  I think what may work is another XML
file that will identify the instances where alternate behaviour may occur as
well as the tests that
will be affected by it.  The rational behind including the "set_switches"
attribute is more for the committee.  Having such an attribute set will aid
use building that "toggle/switches" file.

Greetings,
Carmelo