In various places in schema part 2, we see statements like:

For compatibility (see Terminology (=A71.4)) NMTOKEN should be used only =
on
attributes.

The terminology section says:

[Definition:]   for compatibility
    A feature of this specification included solely to ensure that =
schemas
which use this feature remain compatible with [XML 1.0 (Second Edition)] =


My schema processor has been diligently producing warnings when users =
ignore
this advice, but I've come to the conclusion that these warnings (and =
the
advice itself) are completely useless. A schema that uses NMTOKEN as an
element type is no more or less incompatible with XML 1.0 than one that =
uses
xs:integer as an element type.

The definition itself seems nonsense when you read it carefully. What
actually seems to have happened is that XML Schema includes some =
features
which are provided only to make life easier for people moving forward =
from
DTDs. These features are provided "for compatibility". Someone seems to =
have
then made a false analogy with the use of the phrase "for compatibility" =
in
the XML 1.0 specification, where it is used to advise people how to =
ensure
that their XML is also valid SGML. 

I can't see any logical reason, connected with compatibility or with
anything else, why types such as NMTOKEN should be used only on =
attributes.

Michael Kay
http://www.saxonica.com/ 



From K.Buchcik@4... Tue Nov 22 10:00:20 2005
Received: from maggie.w3