- From: Patrick Stickler <patrick.stickler@nokia.com>
- Date: Thu, 15 Apr 2004 17:24:34 +0300
- To: "ext Chris Bizer" <bizer@gmx.de>
- Cc: "www-archive@w3.org" <www-archive@w3.org>, phayes@ihmc.us, ext Chris Bizer <chris@bizer.de>, Jeremy Carroll <jjc@hplb.hpl.hp.com>
On Apr 15, 2004, at 16:49, ext Chris Bizer wrote: > > I think we should move to OWL in order to be able to express the > cardinality > restrictions, or include them at least into the RDFS comments. I would be happy seeing them expressed in the RDF. Earlier versions had them, but apparently I was doing something wrong so I took them out until Jeremy could say how to define them correctly. > <rdfs:Class rdf:about="&swp;/Warrant"> > <rdfs:label>Warrant</rdfs:label> > <rdfs:comment> > A relationship between an authority and > > ++++++++++++++++++++ > a graph => one or more graphs > ++++++++++++++++++++ I don't read the original wording as equating to a 1:1 relationship. Saying "one or more" seems more confusing, as if the multiple graphs participated in a three-way or more relationship, which isn't the case. > > , in which the authority is in > some way an origin of > +++++++++++++++++++ > that graph. => the graphs > +++++++++++++++++++ Same as above. > Warrants may include a digital signature of > the graph by the authority. > </rdfs:comment> > </rdfs:Class> > > <rdfs:Class rdf:about="&swp;/SignatureMethod"> > <rdfs:label>Signature Method</rdfs:label> > <rdfs:comment> > A method used to construct a signature used to authenticate a graph. > </rdfs:comment> > </rdfs:Class> > > <rdf:Property rdf:about="&swp;/assertedBy"> > <rdfs:label>asserted by</rdfs:label> > <rdfs:comment> > The subject graph > ++++++++++++++++++++++ > originates from and => delete, because asserting says nothing about > who > defined the original graph. It is possible that a graph originates from > Patrick and I assert it adittionally. > ++++++++++++++++++++++ Hmmm... I'm not sure I agree. If you say something, then the utterance originates with you, no matter how many other folks have said the same thing. I.e. to assert a graph is to express the statements in that graph -- and as such, that graph originates from you (and possibly many others). If you make an swp:assertedBy statement about a graph in some other graph, the end result is exactly the same as if you had included that swp:assertedBy statement (and warrant) in the graph itself. No? > is asserted by the authority specified > for the object warrant. The statements expressed in the graph are > taken to > be > claims made by that authority. > </rdfs:comment> > <rdfs:domain rdf:resource="&rdfg;/Graph"/> > <rdfs:range rdf:resource="&swp;/Warrant"/> > </rdf:Property> > > <rdf:Property rdf:about="&swp;/quotedBy"> > <rdfs:label>quoted by</rdfs:label> > <rdfs:comment> > ++++++++++++ > The subject graph originates from the authority specified for the > object > warrant. > Same comment as above. > ++++++++++++ The whole point of making these statements is to associate some source/authority with the graphs. If you quote a graph you are in essences uttering the statements of that graph (even if not making claims by those statements) and, again, whether the swp:quotedBy statement occurs in or out of that graph has no affect on its meaning. To quote a graph is to express/publish/present the statements of that graph. > The statements expressed in the graph are not taken to be claims made > by > that > authority, insofar as any statement using this property is concerned. > </rdfs:comment> > <rdfs:domain rdf:resource="&rdfg;/Graph"/> > <rdfs:range rdf:resource="&swp;/Warrant"/> > </rdf:Property> > > <rdf:Property rdf:about="&swp;/authority"> > <rdfs:label>authority</rdfs:label> > <rdfs:comment> > The object authority is the origin of the graph with which the subject > warrant > is associated. > </rdfs:comment> > <rdfs:domain rdf:resource="&swp;/Warrant"/> > <rdfs:range rdf:resource="&swp;/Authority"/> > </rdf:Property> > > <rdf:Property rdf:about="&swp;/signature"> > <rdfs:label>signature</rdfs:label> > <rdfs:comment> > The object is the signature to be used to authenticate > ++++++++++++ > the graph => the graphs > ++++++++++++ There is only one subject, and hence only one graph in focus insofar as this property is concerned. This does not preclude the same warrant being associated with more than one graph, but the semantics of the terms focus on a specific, single graph. > with which the > subject warrant is associated. > </rdfs:comment> > <rdfs:domain rdf:resource="&swp;/Warrant"/> > <rdfs:range rdf:resource="&xsd;#base64Binary"/> > </rdf:Property> > > <rdf:Property rdf:about="&swp;/signatureMethod"> > <rdfs:label>signature method</rdfs:label> > <rdfs:comment> > The object is the signature method by which the signature specified > for the > warrant subject was constructed. > </rdfs:comment> > <rdfs:domain rdf:resource="&swp;/Warrant"/> > <rdfs:range rdf:resource="&swp;/SignatureMethod"/> > </rdf:Property> > > <rdf:Property rdf:about="&swp;/certificate"> > <rdfs:label>certificate</rdfs:label> > <rdfs:comment> > The object is a certificate by which the authority specified for the > warrant > can be authenticated. > </rdfs:comment> > +++++++++++++++++ > <rdfs:domain rdf:resource="&swp;/Warrant"/> > Certificates should be linked to authorities (or optionally authority > and/or > warrant) They are indirectly linked to the authority because only one authority is allowed for a given warrant. If multiple authorities sign the same graph, then there must be mulitple warrants, one for each authority. So associating the certificate with the warrant also associates it unambiguously with a single authority, the authority for that warrant. Right? > +++++++++++++++++ > <rdfs:range rdf:resource="&xsd;#base64Binary"/> > </rdf:Property> > > ++++++++++++++++ > <swp:SignatureMethod rdf:about="&swp;/X509"> > <rdfs:label>X509</rdfs:label> > <rdfs:comment>X509 Signature Method</rdfs:comment> > </swp:SignatureMethod> > > <swp:SignatureMethod rdf:about="&swp;/PGP"> > <rdfs:label>PGP</rdfs:label> > <rdfs:comment>The PGP Signature Method</rdfs:comment> > </swp:SignatureMethod> > > I think we should wait with instance definitions untill we have > defined the > additional vocabulary that allows you to specifiy a method as a > combination > of cantonalization method, hash algo and signature algo. PGP and X509 > both > allow different hash and sig algo combinations. Thus it is a little > unprecise just to say X509. > +++++++++++++++++ Fair enough. I'll remove them. In fact, we may wish to define the signature methods in a different namespace, or even use existing URIs, if any are appropriate. Patrick > </rdf:RDF> > > Chris > > -- > NEU : GMX Internet.FreeDSL > Ab sofort DSL-Tarif ohne Grundgebühr: http://www.gmx.net/info > > -- Patrick Stickler Nokia, Finland patrick.stickler@nokia.com
Received on Thursday, 15 April 2004 10:25:54 UTC