> Can you elaborate on the workflow, Rob? Al's post seemed to imply that
> inserting semantic tools anywhere would be too time-costly to be
> worthwhile.
Briefly, a typical packet must assume no intrinsic dependencies on remote services. Certain applications require split second guaranteed turnaround. If we can arrange for the semantics to be loosely bound to the packets via references, I suspect we'll find lots of other use cases that will benefit as strongly from "ontological thinking" as any other astronomical application.
Rob Received on 2007-09-13Z22:56:02