Re: union of enumerated values for subtypes

The spec says the union of facets, which means you need to apply BOTH facets. The effect,
in case of pattern and enumeration, is the same as applying a single facet that has the
intersection of pattern or enumeration values.


Kongyi

Stanley Guan wrote:

>  3.14.3
>
>    Schema Representation Constraint: Simple Type Restriction (Facets)
>
>     R - which restrict another simple type definition (B)
>     S - R's {facets}
>
>     The {facets} of R are the union of S and the {facets} of B, eliminating duplicates.
>     To eliminate duplicates, when a facet of the same kind occurs in both S and the
>     {facets} of B, the one in the {facts} of B is not included, with the exception
>     of enumeration and pattern facets, for which multiple occurrences with distinct
>     values are allowed.
>
> Is the above description for enumeration wrong?  In the description, it seems to
> me the {facets} of R, a subtype of B, will have more choices (i.e., enumerated
> values) allowed than B does, which is a contradiction to what a subtype means.
>
> Am I right?
>
> Thx,
>
> -Stanley