Re: Publication request for draft-spinosa-urn-lex

[Barry Leiba <[e-mail]>]

OK, I've looked again at your responses to my review, and at the
changes you've made.  I see that you've made some of the minor changes
I suggested, but that you're holding firm on some of the major points,
which you haven't made changes for.  I'll come back to the two of
those that I think are most important:

* You very much want to mandate a mapping between the title of the
document and the URN.  I think this is very much ill-advised.  You
think this provides discoverability, in that someone can look at a
document title and determine the URN without having to look it up.  I
think this will fail miserably in practice, because there are too many
variables and there will be no way to get it right, in general.
Surely it will work sometimes, but it will fail to work often enough
that people will have to use searches to find the URN anyway, and the
mandated mapping will simple run counter to the filing systems that
some jurisdictions will need to use.

So I stand by my suggestion that you make that a suggested mechanism,
and soften the language that mandates it as the only way.

* There's a bunch of other stuff that I suggested that you soften the
language about, making it clear that it's stuff you have found useful,
without making it a mandatory part of how the URNs are done.  Again,
you think it's important to keep it as it is, and I think that's very
much ill-advised.

I don't see how this scheme can be generally useful as a worldwide
"lex" URN namespace with the kinds of mandatory handling you specify
in this document.

If you absolutely insist on leaving it as it is, I will go ahead and
ask someone to be a document shepherd so we can process this document
and its registration... and the document shepherd will weigh in with
his opinion (which I will not influence), so you'll get another view
on this.  After that it would go out for last call to get the IETF's
and the IESG's view, and see where we are.

I'll stress that I think it's valuable to register this and have the
details recorded, so I *want* this to go forward.  I'd just rather see
it go forward in a form that's flexible enough to actually be useful
internationally, while still representing your model for how you'd
like the URNs to be formed.  My view is that if you put it that way,
then anyone who wants to play your way can do so... and that it's best
not to *require* that in order for "urn:lex:" URNs to confirm to this
spec.

Let me know whether you'd like to continue discussing this, or whether
you'd like me to find a shepherd now, and get a new opinion on it.

Barry