W3C home > Mailing lists > Public > uri@w3.org > May 2000

Re: [Fwd: Re: Approval of initial Dublin Core Interoperabiity Qualifiers]

From: Leslie Daigle <leslie@thinkingcat.com>
Date: Fri, 05 May 2000 18:49:17 -0400
To: Ray Denenberg <rden@loc.gov>
Cc: W3C URI List <uri@w3.org>, Norman Paskin <n.paskin@doi.org>, Larry Lannom <llannom@cnri.reston.va.us>, "Sun, Sam X." <ssun@cnri.reston.va.us>
Message-id: <39134FED.4436AEDB@thinkingcat.com>

Well, first, let's look at the general "URI scheme" vs. "URN 
namespace" question.  In terms of simple pragmatics,

	. URI scheme -- defines both a syntactic structure
	  and some mechanisms for handling the URI (e.g., 
	  resolution); in order to use it, software has to
	  be made aware of it (i.e., specifically coded for it, 
	  plugin, configured for proxy)

	. URN namespace -- may or may not define much syntactic
	  structure, may or may not define specific handling rules
	  (if not it can simply leverage off existing URN scheme
	  resolution mechanisms).  If the latter -- no changes
	  to client software needed (once the whole URN scheme
	  is picked up!  which we've already agreed it has not been,
	  in wide deployment)

Handle defines a complete resolution protocol, so CNRI could well
elect to put it forward as a URI scheme.  If they want global
success (which they may not -- if they have a well-defined niche
market), they also have work to do in order to get browsers
to recognize & parse it.  Again, on the assumption that the URN:
scheme is more generic and likely to get implemented, they
might see advantage in registering a URN namespace and leveraging
as much as possible off its resolution structures as a lead-in to
handle servers, using the handle protocol as a final resolution
protocol. (I hope it's clear that I'm only speculating here, to
answer your question, not particularly advocating one or other path to

DOI (hi Norm!), which is currently based on handles, might elect
to register itself as a URN namespace if they want to keep an
arm's length distance between their identifiers and the current
implementation. (I.e., DOI might be based on handles forever, 
but using the URN approach makes it easier to change if that ever
becomes a necessity).  Once again -- speculation for the purposes
of illustration.

As to:

Ray Denenberg wrote:
> let's say they decide to register it as a URN namespace, and then another
> company develops a scheme that similarly meets the URN characteristics but
> registers it as a URI scheme. Won't this cause confusion?

Do you mean -- if some other organization develops something and
decides to call it "hdl:" and register it as a URI scheme?

Let me point out that there is actually human- and process-intervention
in the registration of URI schemes and URN namespaces.  The IETF
is unlikely to pass anything that was such a collision in strings,
for that very reason.

If that's not what you meant, I didn't understand your question -- 
could you please rephrase and I'll try again.



"My body obeys Aristotelian laws of physics."
   -- ThinkingCat

Leslie Daigle
Received on Friday, 5 May 2000 19:02:37 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:25:01 UTC