Re: Algorithm for merging the pattern facets in a base simpleType with a subtype? (UNCLASSIFIED)

Classification: UNCLASSIFIED
>> Therefore, the patterns that apply to "B" are just the patterns
>> contained in "B". Effectively the patterns in "A" may be ignored. Do
>> you agree?
>
> No, sorry, there is nothing in the spec to justify that conclusion.

This logic is only valid in the narrow situation where "B" contains a
pattern that directly and identically conflicts with a pattern from "A".
In that case inheritance is blocked by instantiation.  Otherwise, if
there is a conflict of patterns and if that conflict is not absolute the
result is typically a union that applies the differences from
inheritance as a remainder for attachment to the portion of the pattern
conflicted in "B".

This sounds simple, except that I did not define the terms "difference"
or "conflict".  I am not sure of a situation where the computed
definition of a single pattern instance can become so complex that the
conflict of it versus a pattern result from inheritance could become
ambiguous.  If such a narrow condition is permissible I would not know
the correct answer.  I suspect the occurrence, if any, of such
complexity would occur more often in the wild as a result of extending
the pattern facet to allow multiple regular expressions in a given
facet.  If this is even a valid case it begs the question of what are
the results in such situations when the hierarchical inheritance is vast
in depth allowing for this condition to be inherited onto a separate
instance of conflict.

Austin Cheney, CISSP
http://prettydiff.com/
Classification: UNCLASSIFIED