[foaf-dev] how to declare that foaf+ssl should be used?

Peter Williams pwilliams at rapattoni.com
Sun Jul 19 21:40:05 CEST 2009


> Irrespective of application of foaf+ssl for facebook's access
> control, foaf+ssl is fundamentally a method for (i) discovering, and
> (ii) walking the friends graph and the wot between any resource, and
> any client's webid.

I think if you stick just to that, then you have isolated the element
of trust reasoning in Linked Data of foaf profiles. foaf+ssl does not
specify how to do that. That is a topic of further research that we
are exploring by using foaf+ssl.

Excellent.



[]Perhaps you are trying to much to look at foaf+ssl in terms of XRIs.

Yes, thats true. Im trying to compare and contrast them. Or more, precisely, Im trying to compare and contrast (i) the praxis of "trust reasoning in Linked Data of foaf profiles" with (ii) the praxis of using HXRI urlrefs bearing "trusted resolution" requirements.



Having done the standard compare and contrast, the ultimate goal of the analysis seeks to go beyond contrasting their apparent nature - aiming to characterize the relationship of their underlying validity logics.



The world of HXRI urlrefs has an explicit (and somewhat overly fixed) validity logic, based on the notion of verified names. It is highly applied to discovery and walking trust fabrics associated with naming constructs. However, its not very general. Furthermore, its highly tied to the iterative procedures ISO standardized in X.509 for certification chain processing. At least, though, it exists and is documented at a level that the likes of me can access.



Now, the logical framework of the various W3C-endorsed taxonomies for the linked-data world have a highly formal validity logic - forming the basis of the distributed computation model itself. But, this is far too general for my needs. What I need is a variant - modal validation logic (or more likely class of logics) that allows the semweb to compute what HXRI trusted resolution already computes (when using HXRI's method of iterative calculation).



I'm actually with W3C in the dispute with OASIS XRI TC. I want the analysis to show that we dont need XRI validation logic as the basis of linked data. ( I seen no reason why the walking the web should be limited to the kind of methods used by DNS-like name resolvers, which is all that HXRI's method really does).  But, as a sideeffect, I do want to identify a usable, modal validation logic -- consistent with the rest of the semwebs LOD logical axioms - that is either a basis of the XRI validation logic or is equivalent to it.



Im assuming, from the way that the foaf ontology defines its terms carefully in terms of higher-logics (i.e. the use of owl in membershipclass) and its cooperation with the wot onology, that foaf will be a crucial element of that modal validation logic f- allowing us to to "assert trustworthiness". Being more applied than the validity logic underlying owl, and being much less applied than XRI's validation logic, it will be suitable for expressing the relative-trustworthiness of anything, including names, resource descriptions or any other web-referencible entity.


More information about the foaf-dev mailing list