RE: Part 2 (Datatypes): whiteSpace not fixed

> -----Original Message-----
> From:	Jeff Lowery []
> Sent:	Thursday, December 14, 2000 5:57 PM
> To:	''
> Subject:	Part 2 (Datatypes): whiteSpace not fixed
> According to the text, the whiteSpace facet is dormant (non-overridable)in
> all built-ins except for string and built-ins derived from string. Yet,
> the
> normative schema near the end of the spec does not show these whiteSpace
> restrictions as fixed="true", as I would have expected. Is this an
> oversight
> of the schema authors'?
Thank you for your comment on the XML Schema specification.

The value of whiteSpace does not *need* to be fixed to collapse in non
primitive types other than string in order to enforce the rule that it cann
be "overridden".

The reason that the value does not *need* to be fixed is that there is a
general constraint on all simple type restriction such that the value space
of the derived type must be no larger than the value space of the base
type...thus, if the value of whiteSpace is collapse in the base type, it
cannot be set to anything else in the derived type, because that would
result in a larger value space.

However, to avoid any uncertainly, the schema for schemas now does fix the
value of whiteSpace for all primitive types other than string.


Received on Thursday, 15 February 2001 21:20:49 UTC