[12553] in Commercialization & Privatization of the Internet
About the creation of new 3-letter domain(s)
daemon@ATHENA.MIT.EDU (Barry Shein)
Sat May 21 04:20:37 1994
Date: Fri, 20 May 1994 17:36:14 -0400
From: bzs@world.std.com (Barry Shein)
To: barrett@daisy.ee.und.ac.za
Cc: com-priv@psi.com
In-Reply-To: Alan Barrett's message of Fri, 20 May 1994 12:07:21 +0200 (GMT+0200) <Pine.3.89.9405201108.R248-0100000@newdaisy.ee.und.ac.za>
>Barry Shein said:
>> >Around here, I don't think there's any official standard way of grouping
>> >the digits.
>>
>> Well, then it doesn't matter in this case, so pick any convention
>> which also works with the proposed standard (i.e. don't make the least
>> significant tuple 3 digits.)
>
>The very fact that there's no official standard for grouping digits in a
>phone number strikes me as a difficulty for your scheme. Given a phone
>number, how will anybody know how to choose between the many possible
>domainised forms? For example, does the phone number 1234567 map to the
>domain 4567.123? 67.45.123? something else?
Again, this is a nit based on a rather rare example that is actually a
worse problem with other schemes, one has to look at it that way to be
practical. Remember the .moscow. vs .moskva. vs .msk. vs .mockba.
example? And that objection arises absolutely everywhere.
I find it a little hard to believe that people actually randomly group
digits in the location you describe.
I can believe there is more than one local convention. Maybe two or
three in actuality?
I also believe grouping, in general, is much stronger and more
intuitive with phone numbers than any other system anyone has proposed
(e.g. geographical), or at worst about the same. It would be folly to
believe, for any system based on widespread human practice, that some
less than perfect cases do not exist. All we can ask is whether this
one is at least no worse in its difficult cases, and better than the
rest on the others.
>It seems to me that, if you want to use telephone numbers, then you
>might as well go the whole hog and use stuff like 7.6.5.4.3.2.1,
>as tpc.int did.
Where it exists the strong grouping conventions are useful both to
users and software builders (for example, the three-digit country
code.)
However, other than the 3-digit country code, your suggestion could be
a fallback for very hard cases tho I think there are easier solutions.
There's no particular reason why a domain server's software can't
regroup all the digits if it's sure it knows better. Grouping is only
a human and software convenience. That is, given 0016177399753 (and
knowing the order) one needs no other information other than what you
want to convert that to, 001.617.739.9753 or 001.617.7399753 if that's
more convenient. It's lexical, which means it's basically trivial, no
"deep" information about what the numbers represent is needed, only
that you wish to regroup them according to your own liking.
When you dial a telephone there is no grouping, you just punch in all
the digits (ok sometimes they use pause and wait for next tone but not
to disambiguate the grouping, only to pass it to the next switch or
whatever.)
The digits w/o any consideration of grouping at all will be ultimately
unambiguous, save which order they're to be read in which has been
discussed. The rest is mostly syntactic sugar.
-Barry Shein
Software Tool & Die | bzs@world.std.com | uunet!world!bzs
Purveyors to the Trade | Voice: 617-739-0202 | Login: 617-739-WRLD
