Re: Fwd: I-D ACTION:draft-daigle-uri-std-00.txt

On Thu, Sep 07, 2000 at 02:51:39PM +0100, John Aldridge wrote:
> I'm quite happy with resources being a flexible, extensible concept.  I'm 
> also happy with the name of a resource being a URI.  I understand, and I'm 
> comfortable with the fact that the entity body associated with the resource 
> may change from time to time; and that two distinct resources may, in fact, 
> have the same associated entity body.
> 
> What I'm not happy with is that there is no widespread agreement on how I 
> tell whether two URIs name the same resource or not.

There are really two answers to this and I'm afraid your not going to like
either of them:

1) if you are just limiting yourself to URIs in the general case and 
_no other mechanism_ then you can't know whether or not they name the 'same'
resource, by definition. I.e. in this very constrained case where the only
information you can have is the URI itself, then 'sameness' is determined
by the URI, not the Resource. 

2) if you are not limiting yourself to the general case of all URIs but
are using some protocol or application arena to include some out of band
information then there is an innumerable number of ways to tell whether or 
not two URIs name the same resource. The reason this is the case is that it 
is _all_ subject to the definition you use for 'same'. Equality WRT to 
Resources is not a universally understood function. It is completely dependent 
on which attributes of the Resource you are paying attention to.

Now, at this point we have to be EXTREMELY careful what vocabulary we use.
I've been using a vocabulary from the URI space. There is going to be
a _huge_ and very subtle distinction between the words I'm using and the
words that you are using. I really think we have to be more concrete
about what definitions we use here so I'm going to take a stab at listing
mine:

Resource: any abstract concept which is identified by a URI
URI:      a unique string that is bound in a 1:1 relationship with a Resource
sameness: a state where two or more Resources have some defined subset of
          attributes that exist outside the URI-Resource binding and that are 
          considered equal within the formal system that defines the meaning 
	  of that subset of attributes and which also defines what the concept
	  of 'equal' itself means.
lexical equivalence: the act of comparing two URIs to see if they are 
	  syntactically identical and thus, by the first two definitions,
	  identify the same Resource.
functional equivalence: the act of comparing two subsets of metadata 
	  surrounding two lexically non-equivalent Resources and determining
	  that, for the purposes of the function being undertaken, the
	  Resources can be considered to have the 'sameness' property in
	  the third definition.
URI schemes: a method of structuring the URI string so that identifiers that
	  contain certain well known semantics are grouped together for reasons
	  of interoperability.

These may not be your definitions but now at least you know where I'm 
coming from....

-MM

-- 
--------------------------------------------------------------------------------
Michael Mealling	|      Vote Libertarian!       | www.rwhois.net/michael
Sr. Research Engineer   |   www.ga.lp.org/gwinnett     | ICQ#:         14198821
Network Solutions	|          www.lp.org          |  michaelm@netsol.com

Received on Thursday, 7 September 2000 10:32:25 UTC