Re: Defaulted values and identity_constraint checking

At 2001-01-04 12:37, David Beech wrote:
>As for the intuition, I personally tend to think of attribute
>and element defaults as a front-end shorthand - it's just as though
>you had explicitly written the defaults in place in the instance.
>Are there situations in which that intuition is faulty?

The WG as a group rejected (when? I think it was design issue
210 in the development-issues list) a proposal to specify more
clearly the order in which augmentation and validation occur.
If I may be allowed to quote the description of the resolution
from the issues list:

   RESOLVED 2000-02-04 (Berkeley ftf): to pass over this topic in
   silence (i.e. not to specify that augmentation occurs before
   validation, nor that validation occurs before augmentation,
   nor that the order is not specified, nor that the order is
   immaterial because any order which obeys all the constraints
   expressed in the spec will produce the same results).
   Dissenting: LBNL, W3C. Abstaining: NCR.

   http://www.w3.org/XML/Group/xmlschema-current/issues.html#AugmentationAndValidation

When I began this note, I was intending to say that since we
had decided not to say anything explicit, David's intuition had
jolly well better be correct since otherwise where would we be -- but
now I notice that last bit where the WG also refused to say
explicitly that the order is immaterial, and I infer that we
would, if David's intuition were faulty, be exactly where we are:
with a problem on our hands.

Next week in London, members of the dissenting minority will be
accepting apologies from members of the majority who have seen the
error of their ways.  Evidence of contrition (e.g. a round or two)
will be expected.

-CMSMcQ

Received on Wednesday, 10 January 2001 03:18:50 UTC