W3C home > Mailing lists > Public > public-awwsw@w3.org > February 2011


From: Nathan <nathan@webr3.org>
Date: Wed, 02 Feb 2011 14:45:20 +0000
Message-ID: <4D496E00.4010403@webr3.org>
To: AWWSW TF <public-awwsw@w3.org>
CC: Jonathan Rees <jar@creativecommons.org>, Tim Berners-Lee <timbl@w3.org>
NetworkResource definition:

Nathan wrote:
>  - NetworkResource (always unnamed)

A NetworkResource is an instance of an interface which can be 
interacted with by agents over a network. For the purpose of 
interaction, an interface is described using a set of identifiers. 
Each identifier in the set relates to a temporally varying property of 
the interface which must be known in order to interact with any 
instance of the described interface. At time t, the identifiers set 
must be resolved within their respective scope, if all identifiers in 
the set have been resolved successfully then interaction with an 
instance of the described interface can take place.

For example, the interface to a NetworkResource may be described as:

_:n1 a :NetworkResource;
   :scheme { [] :id "http" };
   :authority {
    [] :host { :id "lessig.org" } ;
       :port { [] :type "TCP"; :number "80" }
   :path { [] :id "/blog/" };
   :query [] .

If all the described identifiers can be resolved correctly at time t, 
then an interaction (like an HTTP GET) can be attempted by an agent.

The same NetworkResource instance may be interacted with several times 
(at time t, t+1, t+2), but it's impossible to determine if it is the 
same instance, as this is hidden by the interface.

I believe that's true, provably so, and is consistent with URI, the 
web, REST, reality etc.

Thus, just like Representation:

  [] a NetworkResource .

Always anonymous and unnamed.


Received on Wednesday, 2 February 2011 14:47:18 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:21:08 UTC