Re: Proposal: 'tag' URIs

On Sat, Apr 28, 2001 at 05:22:14AM -0400, Sandro Hawke wrote:
> You're preaching to the choir saying "URIs can identify anything."

Cool. Sometimes that's a hard one for people to get over....

> The question is what can URIs do that URNs can not.  Since URNs
> supposedly have additional guarantees/constraints, there must be some 
> difference.  But you seem to be arguing that URNs can identify
> anything, too.

Sure. Anything you can assign a URI to you can assign a URN to (after
all, a URN is just a URI scheme).

> If you can have a URN for "the news at the time you read it," as you
> say, then you can certainly have a URN for "the stuff available via
> HTTP port 80 at cnn.com, at the time you ask."   So that means HTTP
> URLs are also URNs.

Nope. URNs have the requirement that once you assign a URN to its
Resource you can never reassign that URN to some other Resource.
Sure, you could assign a URN to whatever the http protocol gave you
on port 80 at cnn.com and that would be a useful thing because then
I could rely on the fact that, no matter what CNN did to cnn.com or
whether or not CNN even existed anymore, whenever I used that
URN I knew that someone hadn't changed the meaning out from under me.

The special thing about URNs is not some limitation on what you can
assign them to, instead its all about assumptions you can make knowing
that a) the name is independent of some network entity that is part of
the identifier and b) that no one will ever use that name to mean something
else latter on.

-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 Monday, 7 May 2001 16:24:43 UTC