On Mon, 16 Oct 1995, Robert Fabian wrote:
> From my vantage point, the problem with all such schemes is
> that the data can only be sensibly understood in a context. And the
> context cannot be practically stored along with all the data.
> The challenge, as I see it, is how to share enough contextual
> understanding that data can pass between people and across time?
I think that natural language interfaces to databases, in the future,
will be able to customize information based on context and the
person accessing the database.
Before this is possible, I think a set of questions designed to
elicit context needs to be implemented into any such interface.
I'm working on designing this sort of interface and implementing it
in Common Lisp, so I'm thinking a lot about these issues.
-- Andrew Moreno Andrew Moreno <firstname.lastname@example.org>