[12599] in Commercialization & Privatization of the Internet
Re: About the creation of new 3-letter domain(s)
daemon@ATHENA.MIT.EDU (Barry Shein)
Tue May 24 02:15:35 1994
Date: Mon, 23 May 1994 21:39:18 -0400
From: bzs@world.std.com (Barry Shein)
To: Stef@nma.com
Cc: com-priv@psi.com
In-Reply-To: <19352.769326932@odin.nma.com> (message from Einar Stefferud on Wed, 18 May 1994 22:55:32 -0700)
>From: Einar Stefferud <Stef@nma.com> [responding to me]
>Some comments on... BZS: "Here are my plusses for using phone numbers:"
>
> 1. There exist well-established standards for them (format etc.)
>
>Yes...
>
> 2. They are a priori familiar to most of the world's population who
> are likely to ever use DNS.
>
>Yes....
>
> 3. There is virtually no internationalization issue. Digits
> are a near-universal character set.
>
>Yes...
>
> 4. Telephone company directory services and every other form of
> existent infrastructure for looking up telephone numbers (CD/ROM,
> phone books, minitel, etc) can be leveraged with this scheme.
>
>But, all those directories are out of date when published, and their
>owners will surely not load our DNS for us from there records, if they
>could, so all you have done is preempt choices for registrants.
"Out of date" means what, 95% accurate? 98%? Not so bad, consider
current alternatives.
And calling 411 or 555-1212 or whatever directory information is for
your area would be up to date (as up to date as you can expect from
anything.)
This would be mostly to answer the question:
Hmm, Einer Stefferud *must* have a domain name for his home
system, I know he lives in Irvine, wonder if I can get it
by dialing 1-714-555-1212...yes!
of course that assumes some universality of adoption of the scheme, or
knowing that the scheme is likely to apply to Einar Stefferud.
It's far from perfect and requires some assumptions but it's a lot
better than what we have today. Particularly if some conventions are
adopted, like companies generally using their main phone number as a
domain name so I know I can call 1-800-555-1212, ask for the number
for "Big Widgets, Inc.", and be fairly confident that sending email to
info@001.800.123.4567 (that is, the number the 1-800 operator gave me)
is 99.9% likely to work (or info@4567.123.800.001, I get into that in
a moment...)
I wasn't thinking of loading DNS from the phone directory services
(tho it might be useful to share this info with them for
administrative purposes.)
I was thinking of end-user information discovery. If I believe I have
your phone number, by any source, I probably have your domain name.
If your phone number is not your domain name it won't be anyone else's
either (or shouldn't be.) So the failure is soft, I get a host
unknown.
But it sure covers what .com is trying to cover today quite
well. Company phone numbers tend to be reasonably stable and easy to
find (1-800-555-1212, hello?)
> 6. It eases a lot of the potential trademark issues. Your phone number
> is your phone number.
>
>Nope! My phone numbers are all owned by The Phone Company! They have
>the rright to cahnge them any time is suits their fancy.
Sure. But the new one is now *your* phone number. I meant it in the
generic sense, it only identifies you, some company can't come along
and say hey, that's too much like *our* domain.
Whatever phone number you are assigned is yours to use.
When a phone number is changed for whatever reason we always have to
inform people so this doesn't add any major new cognitive burdens.
I could imagine some sort of forwarding IP DNS service if people want
that. That is, queries on the old domain return what you would hope
for a while (your IP address.) It's just an alias or cname in the DNS
db for a while.
>If you don't belive me, look at the wonderful infrastructure already
>in place for X.400 Domain names in the Internet;-)...
>All provided by ytour favourite Telephone Company.
I believe you. I just don't see any big problem with this.
I could see the wisdom in coordinating the adoption of a phone number
domain system convention with the phone companies or CCITT or whatever
rather than just hijacking it. But it's not entirely necessary, we
know pretty much the constraints they work with.
I defer to the old "no system will be absolutely perfect, no system
will satisfy every possible objection one can dream up, it can only be
better than the rest" wisdom.
> 8. It leverages a standard that has developed over 100 years of
> practice and that we know is in widespread, global use and appears to
> solve more or less the same problem we are attempting to solve here.
>
>So have many other namign systems been developed over hundreds, even
>thousands of years. If age of development effort is the criteria...
In its defense few of those systems have been designed to solve such a
similar problem.
Addressing every individual phone in the world in a mechanistic way is
not that different from addressing everything that might have an IP
address in the world.
> 1. To be consistent with current DNS practice they need to be used in
> reverse order and some people might find that awkward (at first,
> whether we'd all get used to them soon enough is an open question.)
>
>OK... Lets see now +1 714 842-3711 becomes 1.1.7.3.2.4.8.4.1.7.1.pid
>Note how the grouping has disolved and it becomes a very long
>non-memorable number.
In another message I developed this idea and showed that grouping is
nearly irrelevant (use whatever you like) save one ambiguous case.
Here it is again in a nutshell:
If we insist on the following two rules:
A. Country codes are exactly 3 digits, padded with zeros if needed
(e.g. US=001, UK=044)
and
B. Any phone number which has a least significant grouping (generally
rightmost in the way we write down phone numbers) of 3 digits
coalesces the last two groups.
(e.g. +99 123 456 789 is treated as +99 123 456789)
then we can type them otherwise in either order and any grouping we
like:
+1 714 842-3711:
001.714.842.3711
3711.842.714.001
001.7148423711
are all equivalent.
+99 123 456 789
099.123.456789
456789.123.099
099.123456789
are also all equivalent, you only cannot type: 099.123.456.789 because
that would be ambiguous (is 099 or 789 the country code?)
But that's the only ambiguous case and it's easily solved: Coalesce.
The order is trivially "fixed" in software, lexically; whichever end
has the 3-digit country code determines the order.
Another possible solution is to introduce a pad character. I like
this less but it's trivial and not mutually exclusive:
099.123.456.#789
the # counts in the length but is otherwise ignored, so that last
field "really" has four digits so cannot be a country code, but the #
is ignored for look-up purposes. This means:
#789.456.123.099
is now unambiguous (099 has to be the country code, so this is in "DNS
normal order" with the most general classification on the right.)
I dunno, like I said, both rules could co-exist. It's trivial to the
software.
>Fine... Lets keep what is good about it then. Use numbers, ala
>Christian's proposal, with some form of grouping to aid memorability
>(critically important), but create a new number space to avoid
>collisions with The Phone Company.
Don't you think the huge directory services and namespace management
functions provided by the world's phone companies are worth a moment's
thought before being tossed aside? I do.
Re: zip codes
They're so far from universal that I think it's hopeless. And the USPS
is just as able to change your zip code on a whim. And they don't have
the right granularity. I can have a few phone numbers for my home,
many of us do. Or hundreds for my business site. You don't have
hundreds of zip codes available for your office building, or more than
one for your home.
>Remember, X.400, by design, gave name space monpolis to The
>Administration in each country, and look at how nicely that is
>working.
Since the proposal is that you can just use the phone number you
already have I don't really see any problem unless phone numbers go
out of fashion.
That's what I mean by leverage. X.400 created a NEW (and monopolized)
namespace. This (phone numbers) uses one whose rules we already
understand and has a lot of inertia preventing too radical change
overnight.
-Barry Shein
Software Tool & Die | bzs@world.std.com | uunet!world!bzs
Purveyors to the Trade | Voice: 617-739-0202 | Login: 617-739-WRLD
