lang ID for "*" (any language)

[Pascal Vaillant]

Hello,

> We’ve had a situation arise in which we’ve found it would be useful to have
> a special-use language ID that could mean ‘any language’.  In BCP 47 terms, 
> this is essentially the “*” language range that can be used in matching, 
> but there are contexts in which “*” cannot be used. Having an alpha-3 
> language ID with the same meaning could be a useful alternative for these
> situations.   

Interesting question.

> -          und: this means that a document has linguistic content but the
> identity of the language has not been determined; that’s totally different
> from this
>

Unless it proves acceptable that 'und' should mean, in certain
contexts, "undetermined", and in other contexts, "underspecified".
A practical question could be: is it possible to imagine contexts,
where both meanings would be useful, and would have to be
distinguished.

I tend to imagine that the answer is yes, so as a first reaction
I would favour your proposal.

Another (context-free) argument is of course the need for clarity and
non-ambiguity of ISO-639 tags, which supports the choice of not using
the same tag for "undetermined" and "underspecified".

Best regards,

Pascal Vaillant