Re: defaults

Sorry, I didn't mean to be obscure.

I think that a system that wants to draw a default conclusion should be
allowed to do so but must endorse its conclusion, and I can use the fact
that, e.g., Frank says so, as reason enough to believe it.  Frank doesn't
have to justify it to me.  And it may be that Frank wants to change his mind,
later, or otherwise retract a default.

I can, of course, sign my own default assertions, but if I make defeasible
assertions, I cannot justify them beyond my signature.  (I may have darn good
reasons for believing these defaults, but if they're defaults, they're not
expressible directly in OWL and I can't produce an OWL justification.)

Did that explain what I meant?

Lynn

Frank van Harmelen wrote:

> Lynn Andrea Stein wrote:
>
> > I respectfully suggest that the answer will be in allowing these systems
> > to make signed (as in digital signature) assertions corresponding to
> > these defaults.
>
> Can you elaborate?
>
> Frank.
>    ----
>
> > If this disturbs you -- e.g., if you worry about what happens when one
> > of these systems withdraws a signed assertion because it has found a
> > reason to doubt its default -- then I think that you have just
> > understood one of the fundamental issues of "doing logic on the web".
> >
> > Lynn

Received on Tuesday, 22 January 2002 17:26:45 UTC