[foaf-protocols] semweb .NET, and foaf+ssl

[Peter Williams]

An update:

I've got my semweb implementation skills back up to where  they were 2 years ago - able to host triplestores, local or remote on sparql servers, and perform queries having added reasoners (that manufacture  a particular reasoning scheme from the reasoner-factory by importing the ontology blueprint on the fly from a HTTP URI). I've added RDFS and Euler reasoned factories, so far.

I am struggling with .NET's linq2rdf's imposition of the linq entity mode on RDF; but that because Im going to have to go back to school and learn advanced generic types and yielding. I'll do that over the next week, so I don't have to think at the compiler level (generics = groupings of function pointers, for the class instances). Taking my C++/modula2 era notion of generics and using them in java/C# world is just not working. TO understand how linq abstracts from the triplestore and query into an entityref/entityset model expressed through generics, needs one to think natively - so one can cast one generic into another (when minting a foaf security token say, known to satisfy such as a particular confidentiality policy).

But, I have been able to do I know best how to do - apply normal security practices to secure communictions. The http listener I have (the REST demonstrator in a windows form with foaf+ssl capabilities) has been augmented with the TripleStore and query engine. But, rather than take REST session nature logically, Im doing it literally by applying .NET AppDomains and partial trust models. These rules on the processes and threads enforce session less behavior, and strictly limit persistence (so I cannot cheat). The interesting part here is to consider the role of the server-side SSL3 cache containing the cert-token and WebID, which is just session information for tokens, in drag. In the java world, components can invalidate SSL sessions (causing several hidden token flows to occur between the n parties to a SSL session context, sending crypto signals in all directions); interestingly, one cannot do this in the .NET world though!



-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.foaf-project.org/pipermail/foaf-protocols/attachments/20091222/cd929f47/attachment-0001.htm