RE: [metaDataInURI-31]: Initial draft finding for public review/ comme nt.

Hi Roy,

Thanks for the comments.

> Please do not confuse identification with identity.  The fact 
> is that any object, when placed in a referring function that 
> results in an N-to-1 answer, becomes an identifier for that 
> 1.  That is what happens when Mark uses a mailto URI to 
> identify himself -- he is using the context of the referral 
> to make the indirect identification by way of a unique 
> identifier that does have an N:1 relationship to him, even 
> though it does not match him directly.  Likewise for VINs, 
> SSNs, etc. People do it all the time, on and off the web.
> 
> How an identifier is created/minted and what it means when it 
> is used are two entirely different things.

I think I understand those things (although of course you are free to tell
me I don't). 

Personally, I am ok with a position that allows context of use to affect
what is identified when a reference is made using an identifier.

I think there are folks who want, or believe, that what a URI identifies is
context independent (or the same in all contexts). I thought that there was
a bias toward that POV in webarch. However, on a quick scan of the
"Identification and Resources" section I don't it is stated explicitly. 

If the resource referenced by a URI varies with context of use (which I
think is what you are saying) we should state that very clearly in Web Arch.

Stuart
--

> -----Original Message-----
> From: Roy T. Fielding [mailto:fielding@apache.org] 
> Sent: 14 July 2003 13:20
> To: Williams, Stuart
> Cc: www-tag@w3.org
> Subject: Re: [metaDataInURI-31]: Initial draft finding for 
> public review/ comme nt.
> 
> 
> Please do not confuse identification with identity.  The fact 
> is that any object, when placed in a referring function that 
> results in an N-to-1 answer, becomes an identifier for that 
> 1.  That is what happens when Mark uses a mailto URI to 
> identify himself -- he is using the context of the referral 
> to make the indirect identification by way of a unique 
> identifier that does have an N:1 relationship to him, even 
> though it does not match him directly.  Likewise for VINs, 
> SSNs, etc. People do it all the time, on and off the web.
> 
> How an identifier is created/minted and what it means when it 
> is used are two entirely different things.
> 
> ....Roy

Received on Monday, 14 July 2003 09:37:41 UTC