Re: URI Revised charter

Leslie Daigle (
Wed, 12 Jul 1995 12:08:03 -0400

From: Leslie Daigle <>
Date: Wed, 12 Jul 1995 12:08:03 -0400
Message-Id: <>
Subject: Re: URI Revised charter

[Roy Fielding wrote:]

> OTHER comments on the draft charter revision:
> I don't think this WG should choose any one URN scheme, or any one
> resolution mechanism, and certainly not any agent mechanism.
> What it should do is define an extensible architecture that allows
> *any* URN scheme to be developed, with *any* resolution mechanism,
> such that these things can be usable when they are ready.

What the working group should do, and has not yet succeeded in doing,
is define what a URN scheme is, what a URN resolution scheme is, and
what operations can be expected on either of the above.  That is, 
the basic characteristics that a system must conform to if it is
to be called a URN naming/resolution scheme.

In the interest of seeing URNs adopted, the group should probably also
identify one scheme that does conform to the definitions that are arrived 

> Re: URAs

> Are they uniform? Only one agent format has been proposed, so maybe.

The proposed uniformity is meant to be at the level of architecture,
not implementation.  They may not have achieved that uniformity yet,
but that's why they are up for discussion...

> Are they identifiers? No.

As Mark Madsen pointed out,

} Debatable.  How much indirection is permissible in an identifier?

It depends how you define identifier, or resource.  If you limit 
yourself to single, existing files, no URAs are not identifiers.  URAs
effectively identify a _set_ of resources pertaining to a particular

> Resource discovery is an important topic, but not one that needs
> to be addressed by this WG.  Should there be a WG on discovery? 
> Sure, but that's a separate issue. 

The original charter of the URI-WG suggests that we are supposed to
be exploring the necessary mechanisms to _support_ resource discovery.
So, we shouldn't be writing applications in the URI-WG, but
we should be making sure that the necessary mechanisms are in place for

URAs are not applications.  They are components that will be useful for
resource discovery _and other_ applications.  As the I-D on URAs 
says, we proposed them because we felt they were an integral component
of supporting resource description for Internet applications, and thus
required a standardized approach.



"Life is chaotic...                                    Leslie Daigle
           ... not random."                  
                  -- ThinkingCat                       Montreal, Canada