Re: URI resolver was Re: RDDL and XML Schemas Proposed Recommendation

[Michael Mealling <michael@bailey.dscga.com>]

On Sun, Mar 25, 2001 at 07:43:52PM -0500, Simon St.Laurent wrote:
> At 06:57 PM 3/25/01 -0500, Michael Mealling wrote:
> >The web and the Internet are about interoperability
> >and the only way that happens is if we all agree on some basic
> >assumptions about how things work and stick to them.
> 
> I hate to sound stupid, but could you explain exactly what those 'basic 
> assumptions' are supposed to be with regard to URIs?

That URIs only have a couple of universal concepts:

1) a URI is bound to one and only one Resource (thanks to Simon for
   making this clear in the past when it wasn't). Notice I capitalize 
   Resource to distinguish it from some application specific definition
   of 'resource' (in RDF for example).

2) URIs do not define any universal concept of sameness or equality 
   other than a URI string is equal to itself.

> I have a really hard time with how the 'U' (uniform, which doesn't appear 
> to be true) 

The term Uniform means simply that the rules for a particular scheme or
for URIs in general are not location dependent. A scheme may have
some component that is location dependent (the news scheme for example)
but the rules aren't location dependent...

> or the 'R' (resource, seemingly optional) have anything to do 
> with the 'I'.

And this is the part that was badly defined in 2396. The 'R' is
defined completely in terms of the 'I'. The Resource here is an
abstract concept in ALL cases. A particular scheme can refine these
definitions but they're still the basic assumptions you can rely on
when you find some naked URI just lying about...

> Yes, I know we've been over this, and I've read RFC2396 and related specs 
> for the past three years, but I still don't see much in the way of basic 
> assumptions.  _Please_ assume I'm stupid for the sake of the list, and give 
> us a foundation in clearer terms than the RFCs have provided.  I don't mind 
> using URLs, nor do I mind using identifiers, but I have a damn hard time 
> with really 'getting' URIs beyond the surface level.

1 & 2 above are really it. URIs are fairly simple things. The problems
come when we conflate application specific concepts of equality and
such with URIs themselves that all the problems occur...

This question is exactly what the W3C URI Interest Group will be
clarifying soon. I hope to have the thing done by the end of May...

-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