On Thursday, Aug 14, 2003, at 18:41 Europe/Berlin, Bullard, Claude L
(Len) wrote:
> Isn't that why some have suggested dereferencing the namespace
> to RDF or RDDL documents? In other words, there are those
> pursuing solutions to this. The fact that the name might
> serve a dual role (lexical disambiguation and named
> location) seems to work. Why not?
>
because context matters. not just what's at the root or what's in the
box.
it's the same problem as delegated authentication. it just has more
levels.
> len
>
>
> From: james anderson [mailto:]
>
> naming the set of symbols is meaningless. one has to name the
> combinations. by which i do not mean the sequencing, dominance, and
> lexical constraints one can express in a document definition.
>
> a system which a-priori names the combinations according to the name
> used to disambiguate one of the respective lexical tokens is going to
> be a dead end.
>
> what does one do with versions? with variations in authority? what does
> one call it when it's a soap wrapper only, and a particular payload
> only, and in combination only? etc.
>
