Re: [urn] URN:META namespace registration request

["Hakala, Juha E" <[e-mail]>]

Hello Dale; all, 

I have been busy all week, and on skiing holiday next week, so I will not be able to respond to your comments and edit the registration request before the last week of February. Sorry! 

As regards terms and definitions, we have not had such section in previous registration requests, but I can of course provide descriptions to novel terms listed below. 

All the best, 

Juha

-----Alkuperäinen viesti-----
Lähettäjä: Dale R. Worley <[e-mail]> 
Lähetetty: lauantai 8. helmikuuta 2020 6.04
Vastaanottaja: Hakala, Juha E <[e-mail]>
Kopio: [e-mail]; [e-mail]; [e-mail]; [e-mail]; [e-mail]; Suominen, Osma I <[e-mail]>
Aihe: Re: [urn] URN:META namespace registration request

Comments on URN:META registration request, version 1, 2020-02-07

This document seems to be an outline or advocacy proposal for the use of URNs.  And it seems to be directed entirely to the library bibliographic community.  As such, it doesn't seem to be designed to be understandable by the IETF community.  And it isn't organized as a namespace registration but rather as an exposition or advocacy.  It really needs to be revised carefully.

Some specific issues are:

The term "metadata" is used without definition.  If I understand correctly, this is used to mean "metadata in library bibliographic records", which is a small subset of the information in the world described as "metadata".  It would help if this was clarified.

The phrase "cool URI" is used a number of times.  Is this a technical term or some sort of commentary?

"PURL" seems to be a technical reference to a specific system.  It would help if this was properly footnoted.

"DCMI community" is used without definition.

The ABNF rule for meta-nss has a component nbn-string, which is not defined.  It appears to be intended to be meta-string (or vice-versa).

There is this text:

    However, URNs should replace "cool" URIs, since using them as
    identifiers for metadata elements is not ideal. There are several
    reasons for this:

    1. It is not possible to provide links to alternative versions
    (e.g. human readable / machine readable) versions of element
    descriptions with one URL.

However, given that "URN" and "URL" are disjoint, and "URL" is a strict subset of "URI", it's difficult to figure out what this might mean, as e.g. it seems to imply that "cool URIs" are all URLs, and that somehow URNs would replace the cool URIs as URLs...

There is this text:

    Rules for lexical equivalence:

    None.

However, defining lexical equivalance is a requirement.  And indeed, the ABNF is annotated with rules for lexical equivalence, "; The entire prefix is case insensitive."

There is a long discussion of resolution processes without giving anywhere near enough information that one could write a resolver client.  I suspect that the document is implicitly referring to a style of resolution system that is common in the library bibliographic world.  But if this document is truly to define resolution for a namespace, it has to actually define it.

The format-code part of the URN is actually a sub-namespace identifier, and it references a particular metadata system which is maintained by a particular maintenance organization.  The text says "As assigned by the respective format maintenance agency", but that implies that codes are self-assigned (a recipe for chaos), whereas the text later explains "New codes are added by the National Library of Finland on request."

Dale