[
Lists Home |
Date Index |
Thread Index
]
Jeff Lowery wrote:
> Uhhh, no. You can't mandate intent through syntax.
Preaching to the choir. I argue that constantly, often to little effect. My
point, though, was that the ostensible reason for which we are exploring a
namespacing scheme is our hope of disambiguating names so that processors may
handle instance names and the content associate with them in a sensible and
appropriate way.
> No semantics, no semantics, no semantics. If I say semantic and registry in
> the same sentence, flame me but good. Starting now. :-)
Great. But, again, what we are trying to do is enable processors to do something
sensible with names, and the associated content, which they encounter in XML
instances. The choice of what to do is the effective assignment of semantics
to--or better the elaboration of semantics from--the particular syntax which we
use to differentiate the instances. Try as you might you cannot take the
semantics--in the form of the differing processing which different syntactic
labels will signal--out of any such registry. Simply understood, such a registry
is a shorthand for the semantics which will be elaborated from processing to be
chosen on the basis of the syntax registered. It is precisely for that reason
that I am arguing against a registry, or at least against a non-local one. On
the other hand, essential knowledge for any local processor to do its own job is
the history of what has worked for it in the past, indexed by the instances to
which any such processing was applied, to compare against the instances which
that processor will encounter in the future.
Respectfully,
Walter Perry
|
