W3C home > Mailing lists > Public > www-archive@w3.org > March 2004

Re: Proposed draft RDF Graph vocabulary

From: Patrick Stickler <patrick.stickler@nokia.com>
Date: Wed, 24 Mar 2004 15:46:54 +0200
Message-Id: <B6339F85-7D99-11D8-858C-000A95EAFCEA@nokia.com>
Cc: <www-archive@w3.org>, "ext Jeremy Carroll" <jjc@hplb.hpl.hp.com>, "Pat Hayes" <phayes@ihmc.us>
To: "ext Chris Bizer" <chris@bizer.de>


On Mar 24, 2004, at 14:18, ext Chris Bizer wrote:

>
>>
>> 1. There are 'agents' or 'authorities' which can be identified by
>> URIs and which can be identified in some way as the owner of a graph.
>> That is, a graph (graph instance?) may be  'owned' by an authority.
>> Exactly how ownership is determined and checked is left open for now,
>> but we anticipate some kind of signature mechanism such as ....
>>
>> 2. There is a class of 'web performatives' which are actions
>> connected with an RDF graph which can be done by agents. These
>> include asserting and quoting. We do not mean to close off this
>> class, and expect that people will invent others. Each one has a
>> recognized rdfg: name which identifies it.
>>
>
> I basically agree, but would like to repeat my comments about 
> "ownership"
> and "authority".
>
> It appears to me that we are having three classes:
>
> 1. Graphs
> 2. Web performatives
> 3. Agents
>
> that are getting related.
>
> There are two variations of graph:
> 1.1 an abstract graph like the sentence "The water is green."
> 1.2.an concrete graph / graph instance, that *is named* and can be
> published. There is a 1-to-many relationship between the abstract 
> graph "The
> water is green." and it's named instances. So I would prefer to speak 
> about
> *graph instances* in the context of named graphs and web publishing.
>
> Web performatives connect *named graph instances* with named agents.
>
> There is also a 1-to-many relationship between a graph instance and the
> agents connected to it via the performative. Thus I'm still not 
> convinced
> that we have to pick one of these agents and declare him as the owner 
> of the
> instance.

There can be more than one first-party agents associated. The
distinction is really between first-party (named in the graph
itself) and third-party (named in some other graph). E.g. given

:G ( :G swp:assertedBy ex:Bob . :G swp:assertedBy ex:John . ...)
:H ( :H swp:assertedBy ex:Mary . :G swp:assertedBy ex:Mary . ...)

then Bob and John have a first-party relationship to G
but Mary has a third-party relationship to G, yet all three
assert graph G.

We can treat the first-party agents as the owners/publishers of the
graph, and the third-party agents as supporters/affirmers of the graph.

> What do we gain in a scenario where the "owner" is the agent who
> did the original naming of the graph (that is how I understand your 
> concept
> of owner) and 20 other agents also perform performatives on this named 
> graph
> instance? I think nothing.

There's a significant benefit to the distinction. Those authorities 
named
in the graph itself, such that the signature(s) validate have a tighter
relationship to the graph, and hence a stronger claim of 
ownership/authorship.

I.e. the owner of the graph is the first-party authority identified in
the graph. Others may assert/affirm that graph, but they are not the
publishers of it -- if they were, they *they* would be named explicitly
in the graph and not someone else.

Basing ownership on who named a graph I think would be difficult to
do in practice. Tell me quick, in ten seconds, who minted the following
URI:  http://sjaiu.bib.to/qijeoaafoail/72812784/gkai.ais ???

Better to have the authority explicitly named in the signed graph, and
treat the first-party authority as the owner/publisher.

>
> Using signatures also don't make signing agents special (=owners), 
> because
> several agent can sign the same named graph instance.

True, and then they are joint owners/publishers, if that signing occurs
in the graph itself.

>
> So wouldn't it be an idea to forget about ownership and authority and 
> just
> use the term *information provider* for agents instead. In our 
> scenario we
> are  having two basic roles an agent can embody: Information provider 
> and
> information consumers. Thus concerning publishing information provider 
> would
> be more precise than just calling the think "agent".

No, I think we need to be clear about first-party vs. third-party
authority. Otherwise, you're left hanging in the wind when it comes
to resolving mutual conflicting claims of ownership all made within
graphs other than the graph in question. Having the authority explicitly
stated in the signed/authenticated/validate graph makes it crystal
clear who the owner/publisher really is, since no'one else would be
able to produce such a graph and have the signatures check out.

Patrick

--

Patrick Stickler
Nokia, Finland
patrick.stickler@nokia.com
Received on Wednesday, 24 March 2004 09:37:27 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:32:25 UTC