mailto: semantics, batting averages, and semantic interoperability

Various people have asked the question:

  Does mailto:foo@example identify an Internet mailbox?

I don't expect semantic interoperability to depend
on a single answer to that question. I'd like to
refine the question. First, let's write it more
formally:

  @prefix : <#>.
  <mailto:foo@example> a :Mailbox.

(using http://esw.w3.org/topic/NotationThree)

Now... is that a logical necessity? No, it's not a conclusion
you can justify with no premises at all. (neither is 1+1=2,
keep in mind).

But consider this argument:

  1. { ?X.log:uri str:startsWith "mailto:" } => { ?X a :Mailbox }.
  2. <mailto:foo@example> log:uri "mailto:foo@example".
  3. "mailto:foo@example" str:startsWith "mailto:"
  4. <mailto:foo@example> a :Mailbox.

I expect semantic interoperability to revolve around
machines checking arguments like that. These machines
will have a trusted computing base with core logic
and a standard library of stuff like string manipulation,
and a policy component.

For example, implements w3.org access control policies would
probably be configured to accept a lot of stuff on the
basis of the IANA URI scheme registry.

The W3C webmaster might see

   A mailto URL designates an "internet resource", which is the mailbox
   specified in the address.
   -- section 3. Semantics and operations of
   http://ftp.ics.uci.edu/pub/ietf/uri/rfc2368.txt

and put

  { ?X.log:uri str:startsWith "mailto:" } => { ?X a :Mailbox }

in a config file. That might get tedious after a while,
and they might use a more sophisticated rule to delegate
the formal transcription, ala "if the IANA registry cites ?SPEC and
the webmaster's key signs { ?SPEC :formalization ?KB }
then ?KB".

Either way, the w3.org access control machine is going to
accept the above proof that yes, <mailto:foo@example> a :Mailbox.

Now that scenario is complicated by the fact that
  - IANA, IETF, and W3C have a special place in the trust network
  - the IANA registry and RFC 2368 are written in plain text,
	not a formal notation

So let's consider Joe Baseball card trader...
darn; that'll have to wait till another time...
gotta go.

-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/

Received on Tuesday, 23 September 2003 19:08:00 UTC