Re: Reasoning over Web Data was: Terminology Question concerning Web Architecture and Linked Data

Adrian Walker wrote:
> Hi Jeremy --
> 
> Adrian Walker wrote:
>  > Absolutely!  But would someone bet their business, or the outcome of a
>  > military operation, on heuristics?
> 
> Jeremy wrote: Yes
> 

I have difficulty in thinking of a single occurrence of either a 
business bet or a military operation that does not depend on heuristics 
of some sort, i.e. guess work, typically by a human agent.

The named graphs approach to trust, that Chris champions, regards trust 
as first and foremost modelled by trust in our human everyday relationships.

I think I will turn your misguided rhetorical question around:

would someone bet their business, or the outcome of a
military operation, on a formal proof of any mathematical result?

or

when has anyone bet their business, or the outcome of a
military operation, except on the basis of guesswork and hunch?

The answer to the former question is, of course not! A mathematical 
proof gives you results about a formal model; but what matters in a 
business bet or a military operation is not the formal model, but the 
'real' world[*]. A formal proof may participate in decision making, but 
the key decision of relevance and actionability is based on hunch and 
intuition.

While formal proof may participate in SW trust, it is only one part of 
the picture - much more important is trustworthiness in the sense of: "I 
trust the BBC to report the news in a fair and balanced way", which 
isn't, and cannot be, subject to formal proof.

Jeremy


* Scare quotes only because I am dubious of the utility of the concept 
of the real world - not relevant to this thread.



-- 
Hewlett-Packard Limited
registered Office: Cain Road, Bracknell, Berks RG12 1HN
Registered No: 690597 England

Received on Wednesday, 1 August 2007 09:15:05 UTC