On 30/12/2021 14:59, Gorman, Pierce wrote:
Richard Shockey once observed that telephone numbers as a namespace
are a form of global language. The entire planet knows how to
use them and they do not rely on international character sets.
<tp>
Well sort of. Working for an international company I got in the habit
of writing
+(44)
in front of the rest of the number which is the recommende format, by
the ITU-T I recall.
When numbers became eleven digits, my local phone company recommended
breaking them up 5 6 or 4 3 4 but did not specify a separator. Some use
space, others other characters such as hyphen-minus (do the French use
comma?) and hyphen-minus has more than one encoding in Unicode (as I
discover when I read the minutes of IETF meetings). I note that your
footing uses period which I think can also be encoded in more than one way.
Most web pages I have used do not understand space or plus or brackets
or anything but sero to nine.
Tom Petch
Whether telephone numbers, or any sequence(s) of integers, I always thought
his observation was profound.
Pierce Gorman
Principal Engineer Sys Arch
Direct 913.439.4368 | pierce.gorman@xxxxxxxxxxxx<mailto:pierce.gorman@xxxxxxxxxxxx>
t-mobile.com<http://www.t-mobile.com/> | Follow T-Mobile on Twitter<https://twitter.com/tmobile>, Facebook<https://www.facebook.com/TMobile> and Instagram<http://instagram.com/tmobile>
[A close up of a logo Description automatically generated]
From: ietf <ietf-bounces@xxxxxxxx> On Behalf Of Vittorio Bertola
Sent: Thursday, December 30, 2021 3:50 AM
To: Phillip Hallam-Baker <phill@xxxxxxxxxxxxxxx>
Cc: IETF-Discussion Discussion <ietf@xxxxxxxx>
Subject: Re: Proposal, open up .arpa
[External]
Il 30/12/2021 03:29 Phillip Hallam-Baker <phill@xxxxxxxxxxxxxxx<mailto:phill@xxxxxxxxxxxxxxx>> ha scritto:
If we were to redo telephone numbers, we would not do exchanges or area codes. They are no longer necessary and to the extent they still have meaning, it is privacy violating meaning.
There would be reasons to keep country codes, though. Both organizational (they naturally federate the namespace and define jurisdictional spaces) and functional (at least you know the part of the world where the person is from, and the languages that are likely to work when communicating).
For the same reasons, some big / federal / multinational countries might want to keep area codes as well. Indeed, even in a mid-sized monolingual country like Italy it is very useful to know at a glance whether a fixed number is from Sicily or from Venice. You could make this optional and subject to user consent, but in many cases it would make sense to get a geolocated number.
We would probably use Base32 instead of just numbers. That gets us to some fairly impressive compact identifiers. 5 bits per character, an 8 character sequence is sufficient for a trillion identifiers.
Sure, from a technical point of view, @KXJL-M4AA is acceptable as an identifier. It is more usable than
@MB5S-R4AJ-3FBT-7NHO-T26Z-2E6Y which gives us 112 bits of work factor.
I didn't look up the science that possibly exists on this, but I suspect that strings made only of numbers or only of letters, even if slightly longer, are easier to remember than mixed strings, especially if there is not a rigid pattern on where the letters and the numbers are. Number-only strings also have the advantage that they are universal, while half of the planet is not native in the Latin script.
It's also true that memorability just for the sake of it - not associated to any vanity value in the string - might be a secondary requirement, now that we have contact books etc. However, it's still useful when you need to add a contact in a non-digital way, i.e. with someone spelling their identifier or writing it down for you.
--
Vittorio Bertola | Head of Policy & Innovation, Open-Xchange
vittorio.bertola@xxxxxxxxxxxxxxxx<mailto:vittorio.bertola@xxxxxxxxxxxxxxxx>
Office @ Via Treviso 12, 10144 Torino, Italy
