New extension for transformed languages

Michael Everson everson at
Mon Mar 5 18:44:01 CET 2012

On 5 Mar 2012, at 17:25, Mark Davis ☕ wrote:

> The first versions of the document had a narrower scope, but during
> the course of development, it was clear that there are a great many
> use cases where people wanted the ability to have a much broader
> scope, and it was widened—not merely to 'translation' but to the
> broader 'transform'. While you may not see the need for those, other
> people did.

It remains the case that the document gives no information as to how "translation" is to be used. Moreover, the term "transform" is so vague as to be nearly useless. It could mean anything.

> However, the system works like language tags; additional subtags can
> be registered to narrow the scope. For example, ru-t-it-m0-ungegn
> means not only a transliteration, but one that follows UNGEGN rules.

Yes, this is an example of transliteration. I already said that a mechanism for identifying transliteration and transcription made sense. 

> And these can also be dated, if necessary to have finer distinctions,
> such as ru-t-it-m0-ungegn-2003.

Yes, yes. It still has nothing to do with "translation" or other "transformations" 

> So if you (or anyone else) wants to,
> you can propose addition subtags, such as one to specify that the
> transform uses another transliteration system, or a broader subtag to
> indicate that it is a transcription (and not a translation or other
> transform).

Here again, no mention of "translation". 

My view remains unchanged. Transliteration and transcription could very well be described unambiguously with -t-. But throwing translation (and other unnamed things) into the mix is reckless, vague, and not an example of good standardization. If you wanted to identify translation you could use -b- (for Babel) or something.

Michael Everson *

More information about the Ietf-languages mailing list