Re: AW: WG: choice of equal attributes

> 
> The way I understand the concept of IDREF, this type contains only
> those values, which are bound by the type ID in the same document.
> Thus, for my problem of a reference to an external element I have 
> to use the more general data type NCName.

Referential semantics is beyond the scope of Relax NG; IDREF are
only checked syntacically, which is, in my opinion is good thing.

Claims are being made that another problem can use the information
in Relax NG grammar to validate referentail semantics, but, again,
in my opinion, first, there is no such program yet, then, it will
be a bad thing, because a grammar should include things it cannot
use (unless they are in separate namespace and just combined
with the grammar, not included).

David