Peter Constable a �crit : > A "language" is an attribute of content, and a "language" ID > is used for > declaration of that attribute. > > A "locale" is an operational mode of software processes, and > a "locale" > ID is used in APIs to set or determine that mode.
Oversimplified, I'm afraid. Consider machine translation software or computer-aided translation tools (e.g. translation memories). In these:
A "language" is an operational mode of software processes, and a "language" ID is used in APIs to set or determine that mode.
I tend to support Peter's interpretation (see his rejoinder).
Your examples both have obvious aspects of content. The translation memory may not be in any particular 'mode', beyond retrieving the data whose attribute is defined by the language tag of interest.
This is very different from 'locale' which really does work like a mode, affecting many types of operations of an application.
I think what you are after is the case where a set of rules (e.g. spelling
rules) are identified by language. However, there seems to me still a difference,
since the applying a spell checker etc. requires data that are in the designated
language, whereas for locale-based formatting, the raw data is usually language
independent.
A./
