RE: off the top issues list

At 01:04 AM 9/10/00 +0100, McBride, Brian wrote:
>graham wrote
> > This is a topic close to my current interests.  I don't think
> > it makes
> > sense to _sign_ a _model_.
> >
> > In saying this, I distinguish between a model, and its
> > serialization.  A
> > model is an abstraction, which captures the essence of some RDF
> > statements.  A signature as applied to some sequence of bits or
> > bytes.  Different serializations of the same model will have
> > different
> > signatures.
>
>Īt is true that different serializations of the same model will
>have different signatures.  I don't follow the logic of why that
>implies that signing the model makes no sense.  If I can sign
>the model, then the same signature will validate it however it
>is serialized and whatever or whoever serializes it.

To sign a model, I think you will need to define some kind of canonical 
serialization, and sign that.

I don't know how one would apply a signature to the sets of abstract things 
called resources, literals, properties and statements.  (c.f. RDF M&S 
section 5, formal model.)

#g
--

P.S. am I the only person who has some difficulty with overloading of the 
term "model" to mean
(a) the abstract, theoretical foundation for any collection of RDF 
statements, and also
(b) some such collection?


------------
Graham Klyne
(GK@ACM.ORG)

Received on Monday, 11 September 2000 09:25:27 UTC