[urn] Re: GS1 URN: submission, version 6.0.0

[[e-mail]]

Version 6.0.0 takes care of most of my concerns, and I am in favor of
approving the registration.  However there are a couple of points that
remain:

The structure of the document might be improved.  Specifically, if
section 2 is the namespace registration, it should work as a stand-alone
document.  However, it isn't stand-alone in one respect:  the Syntax
section gives only an informal grammar, referring to the ABNF in
Appendix A.  ISTM that the registration should contain the proper
grammar.

Now if the whole document was approved as an RFC, then the template
Syntax section could just refer to "Appendix A of RFC xxxx".  But my
expectation is that GS1 wants to get the registration approved faster
than the RFC process.

One possible solution is to move the informal grammar from the section 2
(template) Syntax clause into section 1 (introduction), possibly just
after the listing of the three categories of URNs.  Then move the
grammar from Appendix A into the Syntax clause.  That makes the template
stand-alone but still presents the reader with the informal grammar
before the ABNF.

The second point is verifying that GS1 has, or definitely will, provide
some sort of formal guidance/standard that explains how the URNs are to
be attributed.  E.g., given "urn:gs1:gdd:bie:TradeItem.netContent", is
there text that states what document/text/concept this URN names?  This
is opposed to there being a standardized table named "TradeItem" with an
entry named "netContent" but no explicit text stating what the table
entries imply about GS1 URNs.

Now I expect that for the gdd:bie, gdd:bdt, gdd:cl, and std branches
this will be done, people are unlikely to overlook that.  But I'm
slightly concerned about the <ai> branch.  It appears that <ai> URNs are
semantically collections of attribute names and values.  It wouldn't be
entirely surprising if explicitly specifying that fact (and exactly how
collections are to be interpreted based on their members) might be
overlooked during standards development.

Dale