ID for language-invariant strings

Peter Constable petercon at
Fri Mar 14 01:25:15 CET 2008

> From: John Cowan [mailto:cowan at]

> I think that either is quite defensible, although for "i-default" the
> text must be intelligible to anglophones (as, I suppose, it in fact is).

By no means necessarily so.

> Although these strings are not default *displays*, they are used in a
> situation where locale information is irrelevant (as opposed to
> unknown).
> As for "zxx", the fact that the strings are in natural language is as
> irrelevant as the English origins of "if" or "O_RDONLY".
> I agree that "mis" and "und" are inappropriate.

Your point wrt "zxx" is valid to the extent that the primary, intended *purpose* of the strings is a non-linguistic one. In this application, though, the strings are likely to be actual linguistic strings in some language, so to declare they are non-linguistic doesn't feel quite right.

If I have to choose between the current special-purpose IDs, I guess I might lean somewhat toward "und": the strings are (probably) in a particular language, but *which* language is not declared (and happens not to be particularly relevant for their intended purpose). I'm not totally convinced on that yet, though.


More information about the Ietf-languages mailing list