g'day,
Valdis(_dot_)Kletnieks(_at_)vt(_dot_)edu wrote:
On Mon, 29 Jul 2002 02:11:57 PDT, Einar Stefferud
<Stef(_at_)thor(_dot_)nma(_dot_)com> said:
ORSC was highly successful in this effort, until ICANN chose to
collide with one of our ORSC pre-existing TLDs, and the refusal of
ICANN (and others such as IETF, IAB and ISOC) to recognize this
So what you're saying is that if there's no coordination, there's a problem.
Unfortunately, if you assume the existance of such coordination, what you
have is *one root with multiple registrars*. We already know THAT model
more-or-less works in the ICANN root.
Works? Yeah, maybe for some definition of "works", in that
people can give money to more than one entity to get a
string into .COM, but is that really solving the problem
ordinary users actually want solved? Last time I went to get
a new domain name, I was down to either choosing words from
an Old English dictionary (I went with "gydig", which means
I think I'm either "insane, or touched by the Gods" :-) or
explaining to folks "No, that really is a domain name, even
if it doesn't end in .COM". When I read "works" it sort of
reminds me of that old saw about "the operation was a
success, but the patient died".
I've long suspected that when this debate flares up
periodically and folks get all hot under the collar about
DNS it's to a large degree because the camps are tossing
"apples and hand grenades" over each other's shoulders.
There is much heat and light and fruit salad flying about,
with much impunging of motives and declarations of technical
incompetence, but not much acknowledgement that there might
be an alternate type of oblate spheroid that might be worth
thinking about other than the one you're currently tossing
around.
The traditional IETF view is that the DNS is a stable
mechanism for providing an all-important layer of
indirection between names and IP addresses. In this view,
the DNS is a critical infrastructure component of the
Internet, so providing a single, stable root (and thus a
homogenous, consistent view of the world from anywhere on
the net) is essential. To those who argue for improving the
utility of the DNS as a "directory service", the reply is is
that the DNS is *not* a directory service, so it doesn't
matter that it isn't searchable, or that in particular,
users can't ask for and expect to get any name string they
might want. Fracturing the DNS into multiple services would
clearly degrade its usefulness for its original, intended
purpose, so it should be enough to say "look, the DNS isn't
the service you want it to be, because it's not the service
you think it is. You want a user-friendly directory service?
Go build one, but keep your mits off the DNS, since breaking
it will have consequences you wont acknoweldge or can't
forsee".
I have a lot of sympathy for this point of view. I want my
email to resolve as much as the next person. And I would add
that I certainly don't think anyone is entitled to reap a
.COM-like fortune just because they can set up a couple of
DNS servers and hijack a few million TLD registrations. And
yeah, there are a few folks selling snakeoil involved in
this debate.
But...
An alternative bite of the apple (which obviously gets short
shift in this community) is that like it or not the DNS is
currently being used as a de facto directory service by many
of its users and they expect certain things from it that the
DNS technical community is not providing. In particular and
most critically, believers in this point of view feel
strongly that users should have far more choice in terms of
available names than they currently have, due in large part
to "artificial" scarcity in TLDs.
In this world view, the current state of the DNS namespace -
in which the majority of TLD users live in .COM, and the
vast majority of English (and French and a few other
languages) dictionary strings are already taken - is a
terrible disservice to users and thus is something to be
"repaired". The prevailing view here is that "If it can't be
fixed with the current ICANN-controlled structure then it's
time to fracture the root and open things up to a little
healthy competition".
Now, note that I didn't say that this view is right and the
IETF is wrong, so please holster your weapons for a moment.
I'm also happy to agree, if anyone wants me to, that these
are "political", not technical considerations, although I
don't agree with the coda, which runs something like "and is
thus out of scope for the IETF". I happen to think we're
collectively acting a bit like the guy who allegedly killed
his parents and then argued for leniency because he was an
orphan. We've argued for a while now that the political
issues surrounding DNS should be taken up with ICANN, and
this position has left us with a mess. The IETF has punted
such questions as "how many more TLDs?" and "When?" as out
of scope, and now wonder why those morons don't realize the
harm they might do to DNS in the name of more user choice.
Well, I personally don't like to complain about my users
hammering in screws with a hammer if I can't offer them some
form of screwdriver in its place.
Maybe there really *are* two different debates here. We seem
to be guilty of simultaneously complaining about the threat
inherent in attempts at building multiple roots, while
telling folks to keep their mits off the current root. Now,
this position makes sense only if there's really a single
possible outcome here (keep the current DNS name resolution
functionality with no extensions), but what if what those
other folks are doing really doesn't belong in the current
DNS tree anyways, but is useful to some folks? Shouldn't we
be *encouraging* them to take their work off out of the DNS
tree?
If we agree that the DNS works eminently fine for its
intended purpose, but would be threatened by instability if
we were to allow the sort of
innovation/experimentation/fooling around that some folks
would like to see, then let's pack 'em off to "directory
service land", where people operate additional services to
serve communities of size n, for n less than the whole
Internet. Now, if they did that couldn't we even let them
talk about what hooks might be useful to allow communication
from such services back on planet Internet? This would seem
to bring this debate back to something like "how do we tie
together multiple, sometimes disparate services using a
single communications system"? Hopefully, we can agree that
such a question is in scope for this community, whereas
"when do we get more TLDs" is clearly so far out that it
just makes us tired yelling about it....
- peterd
--
---------------------------------------------------------------------
Peter Deutsch pdeutsch(_at_)gydig(_dot_)com
Gydig Software
That's it for now. Remember to read chapter 11 on the
implications of quantum mechanic theory for time travel
and be prepared to have been here last week to discuss.
---------------------------------------------------------------------