[Foundation-l] Community draft of language proposal policy

[Tim Starling]

Milos Rancic wrote:
> On Fri, Sep 5, 2008 at 1:02 PM, Tim Starling <tstarling at wikimedia.org> wrote:
>> RFC 4646 is a superset of ISO 639. An RFC 4646 language tag typically
>> consists of an ISO 639 code, optionally follwed by hyphen-separated
>> subtags, which identify regional variations (such as en-AU), scripts (such
>> as sr-Latn) or dialects recognised by IANA but not by the ISO 639
>> authority (mostly obsolete with ISO 639-3). It also allows the
>> construction of private-use language subtags, prefixed by "x-".
>>
>> RFC 4646 is the standard in computing, because it's often useful to be as
>> precise as possible when identifying language variations. The original
>> version of this language identification scheme was published in 1995 as
>> RFC 1766.
> 
> And one more question, just to be sure: Internet standards (like
> HTML/XML language tags or so) are using RFC 4646 codes, not ISO 639?

Yes. HTML 4.01 references RFC 1766 (the 1995 version) and XML 1.0
references RFC 3066 (the 2001 version). A W3C article on the subject [1]
states that the most recent RFC in the BCP 47 series is the preferred
standard for XML and HTML documents, currently RFC 4646.

[1] <http://www.w3.org/International/articles/language-tags/>

-- Tim Starling