[12585] in Commercialization & Privatization of the Internet
Re: About the creation of new 3-letter domain(s)
daemon@ATHENA.MIT.EDU (Per Gregers Bilse)
Mon May 23 11:15:05 1994
From: Per Gregers Bilse <bilse@eu.net>
Date: Mon, 23 May 1994 12:58:00 +0200
To: com-priv@psi.com
One thing I don't like about (telephone) numbers as destination
addresses is that they're hard to intuitively verify. Consider an
example from the sidelines:
ftp osi.ncsl.nist.gov
vs
ftam {1,1,1,47:0005:80:005A00:0000:0001:E137:080020079EFC:00}
For interactive use, it's not too big big a deal. If you reach the
wrong host, you'll know. If you call the wrong telephone number,
you'll know. But if you send a letter to the wrong 'email number'
you won't necessarily know; at worst, confidential/private
information leaks out. On the odd occasion I have to send a manual
fax, I always tend to watch that the remote machine identifies itself
with some recognizable company name or address. I hate it if all I
get back is its telephone number -- it tells me nothing I don't
already know.
The lack of intuitive content in pure numbers will inevitably lead to
lots of misrouted mail (leak to neighboring ID because of single
typo) unless the address carries a checksum, or addresses are always
allocated so that a single typo will not identify another ID (similar
to coding schemes for bit-transmissions; to some extent the same as
having a checksum). Consider an example taken from our current domain
naming scheme:
Ireland.EU.net is known locally as ieunet.ie
Italy.EU.net is known locally as iunet.it
I don't think I'm particularly sloppy, but I have on more than one
occasion sent mail to iunet.ie and ieunet.it; it will bounce and I
can spot the typo immediately. Our current naming scheme, by using
more or less meaningful names and words, tends to be leak-resistant
because the address space is very big and near-collisions are
unlikely. It isn't perfect, but it's much better than plain numbers;
I don't relish the thought of having to rely on a numbered scheme
being leak-resistant in the same way.
Directory services don't really solve the problem either; they too
will be prone to typos and misconfigurations, again because numbers
are meaningless and (IMHO) hard to type. A destination address of
jon-smith-big-fat-flyfishing-tandem-dog-atlanta.PVT
is infinitely easier, safer and more comfortable to type than a
globally unique telephone number, a postal code, a social security
number, etc, etc. I hate numbers. And people are humans.
--
bilse <bilse@EU.net> +31 20 592 5109 (dir: 5110); fax +31 20 592 5163
``We used to ! but now we @'' (jensen)
