- From: John Black <JohnBlack@kashori.com>
- Date: Thu, 10 Jan 2008 00:38:38 -0500
- To: "Roy T. Fielding" <fielding@gbiv.com>
- Cc: "W3C TAG" <www-tag@w3.org>
On Mon. Jan. 07, 2008, at 10:55 PM, Roy T. Fielding wrote: > On Jan 7, 2008, at 6:55 PM, John Black wrote: >> But both the abstract trees of Noah's recent imaginings and the >> mathematical abstractions of Pat's original examples are resources that >> can be infinite - or finite but exceeding any "reasonable" limit that >> you set on their representations. So perhaps we should say that any >> finite or reasonably small RDF graph may be an "information resource", >> but that RDF graphs and abstract trees, in general, are not. Sorry, but >> IMO that sounds broken to me. > > There are many infinite representations of infinite resources on the > Web today. The fact that none of us have the patience (or lifetime) > to find out if they would end naturally, as opposed to just assuming > something stupid is going on and hitting the escape key, does not make > them any less representative or the resource any less infinite in nature. My point was not that you couldn't represent an infinite resource, but that you couldn't know that such a resource was in one of N possible states, which is a necessary assumption of Shannon's theory of information, according to Noah, who was trying to use Shannon to illustrate that an RDF graph, or abstract tree, or other mathematical abstraction was an "information resource". If a resource is an infinite loop, spewing out an infinitely long representation of itself, then you cannot *know* that it is in one of (only) N possible states. For no matter how large you set N, the resource may turn out to be in the Nth+1 state. John Black www.kashori.com > ....Roy > > p.s. no, that does not imply all resources can be represented, whether > they be finite or infinite. >
Received on Thursday, 10 January 2008 05:41:00 UTC