[12485] in Commercialization & Privatization of the Internet
About the creation of new 3-letter domain(s)
daemon@ATHENA.MIT.EDU (Barry Shein)
Thu May 19 03:08:15 1994
Date: Thu, 19 May 1994 01:39:12 -0400
From: bzs@world.std.com (Barry Shein)
To: avg@sprint.net
Cc: com-priv@psi.com
In-Reply-To: Vadim Antonov's message of Wed, 18 May 1994 23:39:58 -0400 <199405190339.XAA20425@titan.sprintlink.net>
>From: Vadim Antonov <avg@sprint.net> [responding to moi]
>>So what's the answer, is it .moscow., .moskva, .mockba., all of them,
>>one of them, what?
>
>.msk.su
Ah, but sez who?
Well, this may already be a strong de facto convention for Moscow but
there are a lot of geographical regions below country in this world.
I suppose the underlying question is: Does there need to be one
answer? I suspect there needs to be very few answers, only one might
be harsh but it also might be desireable.
>Does not make any sense to keep full names for often-used domains when
>abbreviated form is fairly recognizeable.
It still begs the question as to what that abbreviated form should be
and by what authority it shall be chosen.
>>Do we insist on transliterating to Latin-1 (or even USASCII)?
>
>YES. It's the only common character set. You want everybody to
>be able to use your address.
For the moment, that's really more a statement on current practice
than possibilities. This is an area that's evolving rapidly.
>>Do we try to eventually encode all of UNICODE into geographic domain
>>names? Do we allow 8-bit characters? 16-bit? More? Do we just hope
>>this issue is never raised or just say "USASCII end of discussion"?
>
>UNICODE is dead-born. It creates more problems than solves.
It was just an example. ISO8859 or whatever, etc.
I mean, you're not saying there will never be any reasonably useable
standard for the various character sets in use, right?
>Essentially,
>UNICODE makes impossible even trivial case-insensitive comparisons
>and does not provide *any* reasonable algorithmic sorting. It's already
>a kiss-of-death on using UNICODE for domain names.
Well, here you're arguing with the example, not the concept or the
problem inherent in the concept. Other than when you say USASCII and
that's that.
I don't completely disagree with your objections to Unicode (tho an
alternate proposal would be useful, the problem doesn't therefore go
away except perhaps for some small application areas that might be
able to say USASCII and that's that. Doesn't work well if you're
typesetting newspapers in Japan.)
>Quite a lot of mid-size PBXes do :-) Also, my home phone number
>was 53-226 for quite some time (in a small city...).
One suggestion I made was to then coalesce that into one 5-digit
tuple, 53226. Or a pad character could be introduced that makes it not
3-digits (I don't know, # for example, 53.#226, #'s are otherwise
ignored except that they count in the number of digits in a tuple).
In fact that's encouraging, I was hoping that any 3-digit example
entailed another short (in this case 2-digit) tuple.
>You should have followed the reference to Borges i provided; his
>conclusion is that there cannot possibly be a global classification which
>is useful and not riddled with exceptions, special cases, questionable
>aggregations and mis-attributions at the same time!
You're not trying to prove that the int'l phone system can't possibly
work are you?
>A mnemonic name is always a simplifyed model (or image) of the real-life
>phenomenon; making names precise requires detailed listing of all aspects
>of the phenomenon, indeed, presenting a copy of the phenomenon (somebody
>offered using DNA sequences already; also see Stanislav Lem's "Sum Of
>Technology" [hidden pun on Thomas of Aquinas :-)] -- "The ideal precise
>model of a cat is the cat itself.").
Yes, and the Kaiser once ordered that a 1:1 map of Germany be
created...until someone pointed out he'd have no place to keep it.
But ultimately we do alright with these more manageable conventions.
You're not likely to have an internet link to each individual cell in
your body in the near future (we can leave that addressing problem to
future generations.)
>Migration is not a problem -- mail forwading and aliasing is easy
>and practically free to keep. I have all my old e-mail accounts still
>useable. Try to do that with telephones :-)
You can do that within limits in some phone exchanges, if you're
willing to keep paying for the old numbers (and any tolls the
automatic forwarding incurs) which few individuals are. Businesses do
it all the time tho.
But hey you work for Sprint that's an RBOC kind of thing :-)
>DNS is popular because it provides an easy way to build easy-to-remember
>and reasonably short unique names. avg@sprint.net is much easier to remember
>than 17039042013 and 18006698303 (which happened to be the telephone
>numbers i can be reached at).
But then again you do deal with phone numbers anyhow, as phone
numbers, right?
String mapping avg@sprint to 001.800.669.8303 is not that hard a thing
to do in a user program.
>X.121 (that's what you're advocating) -- just say no.
Hmm, X.121? What are all those digits supposed to refer to? You expect
me to understand that?!
I can't believe someone from Sprint is arguing against the utility of
phone numbers?! The internet is truly a unique communications medium.
-Barry Shein
Software Tool & Die | bzs@world.std.com | uunet!world!bzs
Purveyors to the Trade | Voice: 617-739-0202 | Login: 617-739-WRLD
