Possibly, the purpose plays some role here: langUsage is for
documenting, xml:lang is there (pimarily ?) for more or less technical
reasons. (Methinks, this could be formulated more precisely...)
langUsage has (should have) 'scope' over xml:lang, too, ideally listing
all the xml:lang values used in the various places in the document,
possibly also restricting their occurrence (though here I may be
following my initiative in FreeDict, of marking the source and the
target language of a dictionary this way).
Still fishing for a clever way of stating this,
On 05.04.2011 21:45, stuart yeates wrote:
> On 06/04/11 05:35, Matthias Einbrodt wrote:
>> As for xml:lang I would say there is a difference, because
>> unlike langUsage and textLang it is not used to denote the
>> language of a document, but to denote the language of the
>> textual contents of an TEI-element.
> I'm sorry, but are you saying that:
> <TEI xml:lang="mi" ...>
> should not be used to denote the primary language of a document? That's
> certainly how we've been using it.