Registering NIDs with purpose of being unique

[Frans Englich <frans.englich at telia.com>]

Hello,

I am informally investigating the use of a URN scheme for an open source 
project, and writes this letter to gather information. I think I have a valid 
use for URNs, but do not know what type of namespace to pick. I hope this is 
the right forum, otherwise feel free to point me in the right direction.

The usage scenario is to identify abstract concepts, error conditions, in a 
globally unique way with URI syntax. In other words, on the syntactical level 
an URL scheme could be used since a domain is to me available, but that would 
be to abuse that mechanism since it is not physical resources that are to be 
identified. The only purpose is to have unique URIs.

Does the scenario qualify for a formal namespace, as per section 3.3 of RFC 
3406? Perhaps you can help me with that question.

The specific URNs composed within this hypothethical namespace would be 
assigned internally, and only have meaning internally. Their use on a 
"public"/"global" level would only be to act as a unique identifier, a 
comparision on the string level. The NSS would never be introspected or 
parsed with semantics in mind.

Let's call the open source project "acme". From purely my perspective, it 
would be perfect if the the NID "acme" was allocated and the NSS was not 
constrained. This would achieve what we only want: the use of unique 
URIs(URNs to be specific), and the project could internally arrange the NSS.

Does this qualify as a formal namespace? Is it a good idea?

I see no direct hinder for being a formal namespace, such as relying on a 
proprietary network, but I neither see it will provide "provide benefit to 
some subset of users on the Internet", except for the only purpose of being a 
unique string.

In essense, the draft would be similar to(but without the domain-name 
complications in this example):
http://lists.verisignlabs.com/pipermail/urn-nid/2002-September/000322.html

How would the reaction be on such a draft? Is my need of a NID valid? If my 
suggestion is nevertheless an abuse of the URN mechanism, what is then the 
proper solution?


Thanks in advance,

		Frans Englich