Lists Home |
Date Index |
On Thursday 24 January 2002 04:36 am, Eric Bohlman wrote:
> As the dictionary entry Len cited makes clear, the
> pre-image of a semantic mapping includes an environment, which
> cannot be contained in the message itself, unless the message
> somehow encodes all of reality in itself, which is not practical and
> IMHO not possible.
Commonly known as the framing problem.
The point here is that "meaning" and "semantics" of XML elements is
really only meaningful in a given application context. My assertion
has always been that in such situations, namespaces don't help
*except* as an alpha-renaming mechanism to aid in disambiguation.
All the hubbub about "semantic discovery" and "downloading semantics"
etc. etc. is nonsense. People have been trying to do this with agents
for years (KQML et al.) and have met with only partial success.
Will it be possible eventually, I think so. Right now, absolutely not.
As such, the trick now is to package the XML, and the interpretation
context for the application (ie. how to package the application).