[foaf-protocols] ACL Append
ojirio at gmail.com
Sat Jun 19 21:46:30 CEST 2010
Right, I was thinking about this a few months ago and I came up with a
way how to do semantic RPC, but let me get back a bit...
So generally a friend request is when Alice wants Bob to put a specific
triple(s) in his profile (FOAF). If we generalize a bit more, it is a
request to put some triples to some RDF graph...
That could be achieved by relating a graph to a resource one can HTTP
POST, PUT or whatever the triples, so for example my FOAF file
http://mud.cz/foaf.rdf would contain triple:
<http://mud.cz/foaf.rdf> todo:putbox <http://mud.cz/foaf.rdf.putbox> .
and the <http://mud.cz/foaf.rdf.putbox> would be the resource for
requesting adding of triples into the graph... (FOAF+SSL authorization...)
This of course has a problem, what about if I didn't want to request
adding of triples, but deleting? Community could then make another
property todo:deletebox for identifying "HTTP inbox" with such semantics...
That of course got me thinking that I might want to request various
things, perhaps thing the community hasn't thought about yet, so I
decided that I want a general "HTTP inbox" where I could send any
procedure defined using some computation/program/event ontology I want
to work on...
(Perhaps best would be just for the HTTP inbox just to accept one URI
instead of RDF data, where I could fetch the data to avoid spamming and DoS)
This is primarily just a vanilla brain dump for me to not forget it, but
feel free to comment and I would be very interested in anything related!
On 06/19/2010 06:15 PM, Nathan wrote:
> Melvin Carvalho wrote:
>> 2010/6/18 Nathan <nathan at webr3.org>
>>> 'If I want to say that I want to be your friend, for example, I could
>>> write that as a simple one-line statement into a "friend requests" file
>>> which you allow me write access to. In fact, I only need append access,
>>> and not even read or general write access to that list.'
>> What would the one line statement equivalent to 'I want to be your friend'
> Following up (rather quickly).. I think we might have the makings of a
> semantic inbox here..
> Given that to Append any information you need to be Identified, thus
> through FOAF+SSL, the triple accepting process will already know the
> data the request was made, and the WebID of the Person who's Appending
> the information.
> Thus this could simply be a generic triple accepting process which can
> be used for virtually anything..
> if the triple is:
> <#you> foaf:knows <#me> .
> then the system or human can read that you know me, your webid can be
> dereferenced to ensure this is true, and then optionally the human can
> (or direct the system to) foaf:knows you back.
> <#you> foaf:knows <#me> .
> then the system or human can read that you have requested i add this
> triple to whatever resource I please, it's a statement you want me to
> make again appropriate action afterwards (if any).
> <#post> foaf:maker <#me> .
> A publishing system has just notified that I have been stated to be the
> creator of something, I can then update the appropriate resource to say
> I made <#post>, or contact the publisher to argue about the fact I
> didn't make the post!
> <#account> sioc:account_of <#me> .
> Just opened up an account in an application, app has pinged back to say
> so, I can update appropriately (or be informed that somebody else is
> trying to open an account in my 'name').
> _:b a x:Message ;
> sioc:content 'how are ya mate!'@en .
> foaf:maker <#you> .
> You just send me a message to my semantic inbox..
> s p o .
> Almost endless scope here - thoughts??
> foaf-protocols mailing list
> foaf-protocols at lists.foaf-project.org
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 262 bytes
Desc: OpenPGP digital signature
Url : http://lists.foaf-project.org/pipermail/foaf-protocols/attachments/20100619/0e692f8d/attachment.pgp
More information about the foaf-protocols