Re: [urn] [IANA #1215575] GDST URN Registration Request Status Inquiry

[[e-mail]]

"Sabrina Tanamal via RT" <[e-mail]> writes:
> We've received the following request from Philip Heggelund. Please see
> below.
>
> It looks like the request was sent to the mailing list back in June,
> but no response.

Below, I recap the situation because many of the relevant messages that
I've seen aren't on the [e-mail] mailing list.  (I don't know why.)
My assessment is that the comments should be considered by the
registrant but that our overall attitude to the registration request is
favorable and reaching consensus should be easy.

Dale

The original request is archived at
https://mailarchive.ietf.org/arch/msg/urn/8V4juiJBSm98NoU-m00Fo2t6TLU/
It was sent 4 June 2020 (not 2021) and was accompanied by an RTF file
containing the registration.

I have seen the following comments about the registration (abbreviated):

From: "Hakala, Juha E" <[e-mail]>
Date: Mon, 22 Jun 2020 06:20:12 +0000

    Having red the introductory material, I am convinced that the
    request is relevant and should be processed ASAP, since an
    international standard using URN:GDST's was approved in March.

    There is no formal specification of the syntax, just several
    examples of URN:GDST's assigned for different resource/data
    types. IMO this is acceptable since clause 6.4.2 in RFC 8141 says:

    [...]

    I cannot tell if what is said about lexical equivalence of these
    URNs is sufficient.

    Organization that is generating the URN is identified with either
    its domain name or with UUID. This is not an ideal solution but
    although domain names do not last for long, they should live long
    enough to enable tracing of a particular product. If the
    organization does not have a domain name and UUID is used, the
    organization can be found using the metadata.

    Data types are not explained, but there is information about some of
    them in the core standard specification. IMO the information
    provided is sufficient.

From: "Lars G. Svensson" <[e-mail]>
Date: Mon, 1 Feb 2021 14:19:17 +0100

    working through my backlog, I managed to process this request today. I
    agree with Juha, that it is a very important namespace!

    I'd say that the use of domain names should be discouraged and that they
    should settle for UUIDs only, since domain names can (and will) be lost.

    Also I miss information if the list of data types is closed and (if not)
    how they will handle additions to that list.

From: [e-mail] (Dale R. Worley)
Date: Wed, 10 Feb 2021 21:43:22 -0500

    The registrant is named as "Global Dialogue on Seafood Traceability" but
    there isn't really any contact information for the GDST itself.  It
    appears from their web site
    https://traceability-dialogue.org/gdst-steering-committee/ that it has a
    Steering Committee, so it's clear that the GDST is a body that can act
    as a whole.  The registration ought to document that.

    A number of the syntactic components are described as "For
    URN-equivalence, the [***foo***] is case sensitive."  That is good.  But
    it would be useful for completeness and accuracy if the [domain/uuid]
    component was described specifically as case-insensitive.

    As others have noted, the syntax specification is unusually informal.
    But as far as I can tell, it is unambiguous, and the registration
    provides useful examples, which make clear what the syntax notation is.

    As others have noted, domain names do not have long-term stability.  One
    way of looking at this is that if a domain name changes hands, the new
    owner could avoid using [prefix] that have been used by former owners.
    (This depends on these URNs either being short-lived "items in commerce"
    or a small and easily discerned set of long-lived "dictionary item".)
    Another approach that has been proposed for other URNs is to combine the
    domain name with its date of acquisition by the current owner.  Another
    approach is to have the "domain names" registered by the user with
    either the GDST or IANA.  But of course, using UUIDs avoids all of these
    problems.

    An environment question that is significant is whether these URNs are
    expected to be seen/read by humans during ordinary usage.  If they're
    almost always contained in barcodes, UUIDs without inconvenience.  (The
    information capacity of QR codes suggests this is easily accommodated.)