[Bug 2664] wd-2: "fixed value constraint" based on identity or equality

http://www.w3.org/Bugs/Public/show_bug.cgi?id=2664





------- Additional Comments From noah_mendelsohn@us.ibm.com  2006-01-05 21:57 -------
I remain somewhat concerned about this resolution, especially given the
possibility that we would add a PSVI property reflecting the value, as opposed
to lexical form, of a validated element or attribute.

As I understand our resolution, it says that I can have a precisionDecimal
element with a fixed="4.000", and an instance with <e>4.0</e>, and the instance
will validate.  What will we report in the PSVI as the value?  If it's fixed at
4.000 it seems odd to report 4.0, since we said the value is fixed and 4.000 is
clearly a different value.  Similarly, it seems odd to report 4.000, since the
instance clearly contains the lexical form of 4.0.

At the very least, I think we should note the possible tie in between this issue
and our possible plans to report values in the PSVI.  

Note that no such concern seems to arise with default="4.000", which has
historically been closely tied to "fixed".  Maybe or maybe not we should take
one more look at enumeration, but it certainly seems odd to have an element
that's explicitly "fixed" and then to report the element has legally having a
different value.

Noah

Received on Thursday, 5 January 2006 21:57:36 UTC