Re: Excess URI schemes considered harmful

At 04:01 PM 9/25/01 -0400, Mark Baker wrote:
> > IMHO, urn:ietf:params:media-type:text-plain
> >
> > is much better....
>
>But what if the IETF cedes control of this registry to some other
>body?  Then that URN would break too.  How is this any different
>than with an URL?

I know I'm out of step with some W3C received wisdom on this, but I believe 
the difference is this:  the urn: form carries a clear and unmistakable 
indication that this name is *intended* to be persistent and permanent, 
usable as a basis for information exchange at any arbitrary time in the future.

I understand the philosophy of "cool URIs don't change", and the fact that 
stability is a social problem rather than a technical problem.  I think 
that having a form of name that carries a clear signal of intent, and whose 
allocation is subject to some degree of consensus process, is a helpful 
element in cementing the social protocols needed to ensure that identifier 
persistence is actually achieved.

Finally, I'll note that I have tried to create a portion of stable URI 
space for persistent identifiers within my company's http: space.  See 
http://id.mimesweeper.com/.  So far, I've sort-of succeeded, but I've no 
great confidence that the identifier persistence will be locked in for all 
time.  Within many company organizations, control of the http: URI space is 
with the web masters, who themselves are part of the product marketing 
group.  The URIs are perceived as simply a way to get to the web pages, and 
are subject to change every time the web site is re-organized or the 
product marketing strategy is reviewed.  In general, these people just 
don't care that a stable URI is a fundamental element of web architecture, 
and will have little patience for some apparently arbitrary rule that 
impedes them from doing their job.

#g


------------
Graham Klyne
GK@NineByNine.org

Received on Tuesday, 25 September 2001 17:14:10 UTC