Re: [foaf-protocols] Bootstrapping the Semantic Inbox

On Thu, 10 Nov 2011 09:38:26 +0100
Henry Story <> wrote:

> Putting on the ping back lenses my first question about the semantic
> inbox, is how does a robot know that it is POSTing to a semantic
> inbox. 

Why would a robot be POSTing to a resource if it didn't know what the
resource was?

You'd tend to know that a resource is a semantic inbox because you
discovered its URI in a place where a semantic inbox is expected.
(e.g. you discovered its URI in an RDF file, as the object of a
statement whose predicate has a range of the class of semantic inboxes)

> Does a GET on the semantic inbox have the resource itself describe
> itself that way?

GET should probably expose the contents of the inbox, usually after an
authentication stage.

I'd expect OPTIONS to provide a description of what methods, media
types, etc are supported by the resource.

As per my original message, a Semantic Inbox is just an extension of
the HTTP Inbox, but the Semantic Inbox does clever stuff (decide for
yourself what stuff is clever enough to qualify) based on the contents
of the messages it receives.

The HTTP Inbox is close in concept to a DAV collection, but while in a
DAV situation, the set of people with write access would typically be
much smaller than the set of people with read access, an HTTP inbox
would usually be the other way around. Write access would be open to
all (except spammers hopefully), while the ability to read the
collection is restricted to the owner of the inbox and his/her

I have, for some time, been refactoring my DataWiki
<> with a hope to release the source code quite

The initial release will not include ACL, but I hope to add
authentication and ACLs soon. Once that's in place, it should
provide a pretty good foundation for a semantic inbox.

Toby A Inkster

Received on Saturday, 12 November 2011 23:05:17 UTC