Re: URx Questions

On 2002-01-23 19:16, "ext Tim Kindberg" <timothy@hpl.hp.com> wrote:

> At 12:21 PM 1/23/2002 +0200, Patrick Stickler wrote:
>> With regards to temporal anchoring, 'tag:' seems to me to be
>> too "mothering" and not general enough.
> 
> You seem to have missed the point of tag. Without date qualification (or
> something much more laborious), an entity that holds an authority name
> could pollute the namespace of another entity that earlier held it. The
> restriction is solely to guard against that. It is not exactly a
> resource-hungry restriction.
> 
> Of course, if you really want to, you could buy some DOIs instead (assuming
> that _they_ will always be around). But we wanted to allow individuals and
> small organisations to participate in naming at little or no cost -- that
> includes those who possess only an email address (or domain name). Email
> addresses are quite frequently given up as users change ISPs so the chances
> of someone else getting the same authority name are not negligible.

I fully understand the motivation and benefit of the tag date qualification,
I just don't see that it has to be manditory. I think I gave sufficient
examples to justify that view.

And if a minting authority is truly concerned with absolute temporal
uniqueness, they can either include date components in their hrn or
use the UUID form.

You seem to argue that because date qualification is not manditory,
it's not possible, or that hrn's can't achieve temporal uniqueness
without it.

> 
>> And as has been pointed
>> out, does not provide a form that garuntees against accidental
>> collision within the same authority (e.g. UUID).
> 
> <repeat my earlier remarks>
> 
> 
>> I think that the IETF should *heavily* discourage, if not disallow
>> dual registration of equivalent URI and URN NID schemes.
> 
> I agree with you. Tag is not where it is by overall design but by where and
> how we thought we could add value at the time. I think it's a mess that we
> have both registrations going but 'the system' is a mess and I'm not yet
> sure what to do to right our situation. I'm hoping that one or other
> registration will reach a stage when the solution to the problem becomes
> clear.

It seems to me that the deciding criteria whether one registers
a URN scheme as a top level URI scheme or a 'urn:' NID is whether
one plans to use DDDS for global, transparent resolution.

If so, then the NID makes sense. If not, then I see no reason
to bother with NID and go with URI registration (though the
latter is rather scary and laborious).

Whether DDDS can be used or extended to support arbitrary URN
schemes will be an interesting exercise.

Patrick

--
               
Patrick Stickler              Phone: +358 50 483 9453
Senior Research Scientist     Fax:   +358 7180 35409
Nokia Research Center         Email: patrick.stickler@nokia.com

Received on Wednesday, 23 January 2002 13:53:32 UTC