[
Lists Home |
Date Index |
Thread Index
]
> The values in the camera document changes as the instance
> of the camera changes. The values of the schema of the
> camera changes when **cameras** evolve. This relationship
> of schema to document, type to instance, can only be evolutionary
> when an observer of camera evolution changes the schema.
Okay, to take another angle it seems that much of the problems would be
solved if we could isolate change. In that case, is this anything more than
a versioning problem (e.g. in time and space/locality)? Suppose that the
only thing that had to be controlled is the area of influence for the system
(as defined in the schema or as defined by the area... e.g. Roger's Camera
store could use the version http://cameras.org#NorthAmerica *or* you could
say any document in NorthAmerica belongs to the aforementioned area.
Personally I prefer the former) Secondarily you must denote the effective
dates useful for given ontology. For example:
term: "bad"
equivalent: "good"
location: http://example.com/slang#UnitedStates
time: 1980-1991
Of course there are trust issues and such... but ignore that for now. If you
could manage these aspects (as you would in any business rules system) you
would be a few steps closer. Also, a reasoner may be able to deduce
assumptions about an overall document based on specific items within the
document to determine locality and time.
I am probably way off track here...
Jeff Rafter
|
