Re: erratum Re: resources and URIs

>Hi Pat,
>
>>  But in any case, if this really is what the URI<->resource connection
>>  is supposed to mean, this ought to be clearly and unambiguously
>>  spelled out in the TAG architecture description.
>
>I hope your continued persistence will drive this discussion to such a
>profitable conclusion, so that we may speak of URIs and resources with
>confidence that we have some shared interpretation of the terms :)

That's my goal.

>
>Just thinking out loud, on your recent examples:
>
>The web is a system for retrieving representations from and transmitting
>representations to things addressed by URI, and those things are called
>resources. The web does not depend on what resources actually *are*, which
>leads to the discussions over whether the resource is the weather, or the
>document describing the weather.

Well, OK, except that its hard to understand how the weather could be 
the transmitter of a representation (isnt it? It is for me...)

>
>To continue your postman example, the postal system is the web, the
>mailing address is the URI, and the thing that sends/receives letters is
>the resource.

And HTTP is the mailman, right.

>The resource could be a person, or a company, or something
>even more virtual, but its precise nature does not matter much to the
>postal system.

Agreed, and the Web analogy.

>Could you describe the architecture of the postal system
>without in fact describing the nature of senders/receivers at all?

Well, maybe the *nature*, yes; but you do I think need to have as it 
were a postman's API defined. For example, it is OK to leave letters 
in a mailbox, but not on the side of the road. I think there are 
actually laws about things like this, in fact.

>
>However the semantic web appears to be concerned more with reasoning about
>the nature of senders and receivers, rather than simply delivery, and I
>think that is not something that can just be slipped into the existing web
>architecture casually.

I honestly think that most of the SW is not going to be concerned 
directly with Webbish matters at all. It mostly going to be about 
things like money, dates, contracts, people, boring everyday business 
stuff. To the extent that it does get involved with senders and 
receives it is going to be terribly strict and fussy about its terms 
and getting them right, but that's just the nature of the beast: 
formal ontologies do get strict and fussy about defining everything.

Pat
-- 
---------------------------------------------------------------------
IHMC	(850)434 8903 or (650)494 3973   home
40 South Alcaniz St.	(850)202 4416   office
Pensacola			(850)202 4440   fax
FL 32501			(850)291 0667    cell
phayes@ihmc.us       http://www.ihmc.us/users/phayes

Received on Tuesday, 22 July 2003 02:34:12 UTC