[Gzz] Re: Address meanings, not contents! (Re: Storm blocks and metadata)

[Benja Fallenstein]

Hi Reto--

We're definitely talking about different things now. With Storm, I want 
a storage system where you can make a reference that does not depend on 
trust to be resolved: You should be able to download a document from any 
place that has a copy, without dependance on trust; and (though this 
isn't fully implemented yet), you should be able to find the newest 
version of a document published by a given agent.

I believe that it is important to be able to read and link to documents 
whose authors you do not trust. I believe that it is important that 
documents are retrievable as long as there is a copy on the network, 
even if you do not trust the peer which holds the copy.

If a given urn-5 identifies the concept 'the Andromeda galaxy,' I don't 
dispute that anybody on the network should be able to associate 
documents about the Andromeda galaxy with this urn-5, and that clients 
should be able to retrieve this resource. But that's not the technical 
problem I'm trying to address; I'm trying to address the question how to 
make a stable reference to one version of one resource, whereever it is 
located, resolvable without placing trust in anybody.

Does that help us to get back on track with the technical discussion 
about this issue? :-)

Thanks,
- Benja

Reto Bachmann-Gmuer wrote:
(snip)
> > > An example to be more explicit:
> > > <urn:urn-5:G7Fj> <DC:title> "Ulisses"
> > > <urn:urn-5:G7Fj> <DC:decription> "bla bli"
(snip)
> > Secondly, I don't see a reason why the URI of the image would need to 
> > refer to this.
>
> me neither ;-). There must be a misunderstanding here.

(snip)

> > Um, using a urn-5 doesn't work since it's just a random number-- if we 
> > use just a random number, we cannot check whether the data we may 
> > retrieve from a p2p network is really what the person making the 
> > reference wanted us to see. We would need to use "urn:foo:ref:[blah]", 
> > which would be the above RDF data, from which we could then get the 
> > specific representation.
>
> The urn-5 URIs are intended to reference a certain 
> concept/idea/meaning/topic, peoples are free to associate attributes to 
> existing URIs. They may be subject to  change  like terms in natural 
> language are, if somebody wants to use a term in a specific sense she 
> has to make this explicit, maybe using digital signature stuff, but more 
> often I think a key free trust system 
> (http://www.w3.org/2002/03/key-free-trust.html) is not only enough, but 
> more adapted to "fuzzy" trust levels in a P2P network.

(snip)

> > > you could server the uri of both the abstractions (urn:urn-5:G7Fj and 
> > > urn:urn-5:lG5d) directly using http 1.1.features.
> >
> > (Again, you'd have to use hashes, or you could be arbitrarily spoofed.)
>
> (Again. No good networking without trust mechanisms ;-)
(snip)