tagging of existing text vs. tagging of constructed text
addison at lab126.com
Thu Dec 20 18:10:36 CET 2012
> Am I right in noticing that the use of discussed "gender speech subtag" actually
> relates to the text/speech construction?
Maybe and maybe not. It might be used to select between alternate resources ("different texts") based on e.g. gender, in which case the text isn't being constructed, it is merely part of normal content negotiation. But it might also be used to tell some process to construct content in a particular language (which might need to include gender information to better personalize the experience).
> Until now I understood the IETF tags system as the meta-information system for
> the existing (already constructed) texts.
Language tags can also be used to identify the language of artificially (runtime) constructed texts (where "texts" is understood to include any language material, written/signed/spoken), or, indeed, any language material. In addition, language ranges (see RFC 4647--the second part of BCP 47) can be used to describe a user's language preferences.
For example, if you get turn-by-turn directions from a mapping program, the messages displayed or read to you will be in a language and can be tagged with a language, even though they are generated dynamically by a computer program and don't exist as static bits of text unto themselves.
More information about the Ietf-languages