Standard list of languages: suboptimal choice

[Stephane Bortzmeyer <bortzmeyer@xxxxxx>]

The proposed text says:

> For language identifiers, ICANN will adopt the International
> Standard Organisation's 639-1 naming system for identifying and
> labeling particular languages:

There are two strange things in this choice. One is that ISO 639-2 and
639-3 are already issued (639-3 is more than one year old) and they
offer a more comprehensive list.

But the other is more important: language codes are not sufficient to
express all the necessary details. The proposed text adds the country
codes but not the scripts, which are far more discriminating. A
mexican can speak es-AR (the spanish spoken in Argentina) but he could
certainly not read spanish written in the arabic alphabet (as it was
in Spain during the muslim rule). Many languages spoken today are
written with different scripts.

There is already an international standard to express all the details,
IETF BCP 47, currently RFC 4646 (available at
<http://www.rfc-editor.org/rfc/rfc4646.txt> and described at
<http://www.langtag.net/>). It is used by many important Internet
protocols or formats (such as XML and HTTP) and the registry for this
standard is managed by... IANA, a function of ICANN. It is quite
suprising that ICANN references Wikipedia instead of the registry it
manages! (The registry is at
<http://www.iana.org/assignments/language-subtag-registry>)

The reason may be that the full power of Internet language tags is not
necessary for the few languages that ICANN will use (currently, there
are more than 500 languages in the Internet language tags registry,
far more than what ICANN could use). Is that so?