xml-dev - Re: XML Schema Question: default namespace misses attributes

[John Cowan <jcowan@reutershealth.com>]

Michael Rossi wrote:

>    Again, this is undoubtedly _technically_ correct. But I can't see how it
> helps in real applications. This should really be very simple: by definition
> (and no, I haven't looked it up word-for-word) an attribute modifies an
> element, it's metadata. It is absolutely, positively, unequivocably attached
> to the element it modifies in every way. I don't see how it makes any sense
> to say that it can be defined in some other namespace. What's the point?

It makes sense when we want to talk about what the Rec calls "global attributes",
though "unattached" might have been a better name.  The "class" attribute
of most HTML elements is in some sense the same attribute everywhere,
and can even be exported to non-HTML elements coherently.  Therefore,
it makes sense to use "html:class" to designate this unattached attribute.

Such attributes belong not to the element in which they appear, but directly
to the namespace in which they are defined.  (Throwing proper terminology
to the winds here.)   However, unattached attributes are distinct from
elements of the same name.

Therefore, if a namespace contains n elements, it consists of n+2 partitions:

one to hold the element names;
one to hold the unattached attribute names;
n to hold the attributes associated with each of the n elements.

	"Converting people to namespaces, one brain at a time."

-- 

Schlingt dreifach einen Kreis vom dies! || John Cowan <jcowan@reutershealth.com>
Schliesst euer Aug vor heiliger Schau,  || http://www.reutershealth.com
Denn er genoss vom Honig-Tau,           || http://www.ccil.org/~cowan
Und trank die Milch vom Paradies.            -- Coleridge (tr. Politzer)