[
Lists Home |
Date Index |
Thread Index
]
Or the entire framework has a root called "object".
Smalltalk, yes. Elegant.
Urrmm... calling the GI a type has descriptive
usefulness, but using it as a type is a higher
level processing implication. Isn't that precisely
why RELAX made validation its focus?
Isn't that why people want a pipelined architecture?
Nothing Steve claims changes the way Jonathan wants
to *use* XML. Both can be right but Steve is arguing
that Jonathan's use is one of many possible as long
as namespaces don't lock the GI to a *type*. Isn't
that why the spec stayed silent on any *use* past
disambiguation?
Parts and assemblies: it's in the way that you use it.
len
-----Original Message-----
From: Sean McGrath [mailto:sean.mcgrath@propylon.com]
[Jonathan Borden]
>What is the myth? "isa" links have been used forever. A "name" is a
>character string, right, what is open to interpretation? But please tell me
>about this "thing" that is different from its "name". I suspect that if you
>can describe it with enough precision that the answer to these issues will
>become apparent. For example do you wish to describe each and every element
>that exists in the entire universe as a distinct "it" that "has-a" set of
>properties. That would be tedious.
And phenomenological to boot.
Mind game: Imaging a UML tool where all classes are instances of a single
uber-class with an attribute called "type".
|