[urn] Re: third-party registrations

On 4/3/26 8:46 AM, Michael Jones wrote:
> Peter, you wrote "3. Personally I see no rationale for third-party registrations. If someone can provide a strong argument for them, then let's discuss."
> 
> GLUE is a perfect example of why third-party registrations are necessary.  It glues (pun intended) together disparate and bespoke 

Somehow I doubt that DUNS, GLN, ISO, LEI, PEN, and SWIFT think of their 
identifier systems as "bespoke".

> identifier spaces that are not URIs by creating URIs for them.  It creates URI representations for these identifiers when the external authority that defined them didn't do so.  This meets a need to facilitate trade and commerce, by allowing these identifiers to be used in protocols.

It seems to me that there are other ways of using identifiers in 
protocols than by creating a third-party registration and thus (as I see 
it) going against the underlying management principles of the URN 
ecosystem. For instance, each of the aforementioned organizations could 
separately register URN namespaces for their identifier systems, and 
those URNs could then be used in the relevant protocol slots.

> For example, if ISO/IEC 6523 had created URI forms for their identifiers, GLUE wouldn't be needed to do so.  But because they didn't, a third-party registration is appropriate and necessary.
> 
> 				-- Mike
> 
> P.S.  Yes, we're proceeding to update the GLUE spec to use glue: rather than urn:glue:, per your suggestions.  But the general point about the appropriateness of and need for third party registrations remains.

We'll have to agree to disagree.

Peter

_______________________________________________
urn mailing list -- [email protected]
To unsubscribe send an email to [email protected]