Lists Home |
Date Index |
Tim Bray wrote:
> BTW, the referenced page has one error: there is no ambiguity at all
> about what namespace an unprefixed attribute is in; it's not in any
> namespace under any circumstances.
> I personally think this was a design error, but we should either fix it
> (unlikely now, though I'd support an effort) or live with it. The
> document's recommendation to "treat unqualified attributes as being in
> the namespace of the element that contains them" is just wrong, will lead
> to poor interoperability, and should not be followed.
But what does it mean, really, for an attribute to be 'in a namespace',
separate from that attribute appearing (or being declared in an ATTLIST as)
within the scope of a particular element? Are there qualities which we are
to imagine that an attribute draws, or inherits, from a namespace which
influence its meaning or its usage in any way comparable to what that
attribute draws from its dependence upon the particular element which it
modifies? And if not, is Simon's suggestion not an appropriate
acknowledgment of the power exerted upon an attribute by the element upon
which it must depend?