--On Friday, July 10, 2015 07:39 -0400 Keith Moore
<moore@xxxxxxxxxxxxxxxxxxxx> wrote:
> "The world" uses URLs. There are, in addition to that,
> communities that use DOIs, and communities that use URNs.
> And relative to the expected useful life of either of the
> latter kind of names, only a tiny amount of time has elapsed.
> These identifiers are intended to last centuries, but have
> been around for less than 20 years. There's only a tiny bit of
> buyin into either kind of name relative to their intended
> scope.
Keith and many others,
I think we are going around in circles. We may not be
converging, but I don't think repeating variations on the same
theme or denouncing each other's positions, is going to help
much. As you have pointed out, the discussion itself has
gotten expensive. I hope that the IAB has whatever information
it needs from this list. If that is not the case, I hope they
can come back with specific questions or other ways to focus the
discussion.
I think there is at least rough agreement that DOIs in RFCs are
here to stay. Some wish they weren't, or that there had been a
different review process that might have led to a different
decision, but the question of how we got here -- including, for
those who believe it is a serious problem, who is accountable
and how-- is really separate from the ones of where we are and
where we are going.
Part of the "how we got here" question does affect the document,
which is what claims it makes about why this is being done.
Some believe that adding DOIs will make RFCs academically more
credible; others consider that to be, at best, an untested
hypothesis. There seems to be some agreement, even among those
who otherwise disagree, that the document should describe what
is being done and minimize the controversial explanation and
claims about why.
Similarly, some believe that use of DOIs in documents without
using URNs sends a powerful message and that either we need to
move to immediately incorporate URNs next to the DOIs or that we
need to document, somewhere, a clear and persuasive reason for
not doing so. Some who don't believe that about URNs or the
relationship fundamentally don't like URNs or believe they have
had their chance and failed. Others are convinced that there
are more URNs assigned and more use-instances of them in various
contexts than there are DOIs. It is probably possible to
evaluate those particular claims only with a lot more agreement
about what and how to count.
In any event, those who don't like URNs, or don't like them for
some specific set of functions, should be, at least IMO,
starting a serious discussion about whether the WG should be
shut down, or an applicability statement produced, or something
else, not complicating this DOI-in-RFCs discussion by using it
as a surrogate. I note that there have been attempts to start
that discussion. They have gone nowhere, perhaps because
starting it with assertions about, e.g., "useless prefix" or
even how other things have won in the marketplace are
sufficiently inflammatory to cause people to disengage or stop
thinking and start ranting. Personal opinion based on
discussions with people very far outside the normal IETF
community: URNs are heavily enough deployed in some communities
that, while the IETF could abandon them or even denounce them,
the effect would be development elsewhere, very possibility
developments that would lead to incompatibilities or
interoperability issues, not making them go away. It seems to
me that is as much a part of our reality as perceptions about
what other technologies are used, and where, and how dominant
they have become.
So I hope we can wind this discussion down. And, again, if the
IAB needs information it doesn't think it has yet, I hope they
will tell us.
best,
john