Re: [urn] Request for urn mrn

[Kasper Nielsen <[e-mail]>]

Hi Lars,

Thank you for your comments. My comments are inline.

I know there are some sections that might be a bit weak. Having this kind
of meta URN namespace is a bit of an experiment for us. And we hope to
gather some valuable feedback once people start to adopt the standard.
Expecting to strengthen the specification as well as provide better
guidelines over time.


> Eventually I also got around to review your request. It looks almost ready
> to me, but there are two areas you might want to look into:
>
> 1) I cannot find anything on how you want to ensure that the identifiers
> (e. g. the URNs) remain unique and that they are not by accident assigned
> twice. In the "Assignment" section you write that "Management of
> Organization-Specific Namespace IDs should be done in    accordance to the
> guidelines provided at http://www.mrnregistry.org." If that enough to
> ensure uniqueness of the that's fine with me but it would be helpful to
> have that explicit in the registration document.
>
I've changed the section to this:

++++++++++++++++++++++++
   Assignment of Organization IDs is done by filling out and sending in the
   registration template, as detailed on http://www.mrnregistry.org. IALA
will
   ensure the uniqueness of Organization IDs by ensuring that an ID can
only be
   assigned once. Records of all successful registrations will be
maintained in
   a public version control repository as detailed on the website.

   The assignment procedures for MRN strings under a particular
   Organization-specific namespace ID is the responsibility of the
maintaining
   organization. The considerations provided by http://www.mrnregistry.org
as
   well as Section 6.4.3 of the URN specification [RFC8141] must be taken
into
   account when defining the actual assignment procedure.
+++++++++++++++++++++++++


> 2) There is no section on lexical equivalence (e. g. how you can decide if
> two character sequences are "the same" URN). The "Syntax" section says:
> "The namespace, “mrn”, is case-insensitive in processing but is
> conventionally written in lower case" but that's about all. Since much of
> the handling is done in sub-namespaces, I understand that it can be
> difficult to give exact rules here, but I think that the registration at
> least should say that for comparison, the strings "urn" and "mrn" must be
> converted to lower case. If you can have such a rule for the OID, too, it
> would be excellent. Beyond that, I guess that it depends on the how the
> organisation identified by the OID handles its OSS.
>

I've added the following section:

+++++++++++++++++++++++++++

Rules for Lexical Equivalence:

   In addition to the URN equivalance rules defined in Section 2.1 of the
URN
   specification [RFC8141]. Case normalization must be applied to both the
OID
   and OSNID part before determining lexical equivalence.

   Rules for equality comparisons of the OSNS part must be clearly
documented
   by the governing organization.
++++++++++++++++++++++++

/Kasper