draft-phillips-langtags-04 /2.3 UND
tex at xencraft.com
Thu Jul 1 22:45:03 CEST 2004
With respect to recommending that UND not be used:
There is a difference perhaps in an unlabeled document (perhaps the authoring
tool doesn't support language tags) and a document for which the language has
not (or could not have) been determined.
For XML, SOAP, and RDF, there was discussion that content embedded in other
content inherits the language of its container.
In that situation, there needs to be a way to prevent inheritance where it is
not appropriate, since the content may be a different language from the
Perhaps the UND bullet should comment on blocking inheritance and whether
omitting the label vs. perhaps giving the label a null string value should be
With respect to language matching, how do I look for documents that are not
assigned language tags?
It seems that UND might be useful for this purpose and is distinct from using
an empty language tag as a match value, which (I think) matches documents in
any language (same as "*").
So: language-range = language-tag / "*" / "UND".
Hmmm. If I read the BNF correctly, language-tags that are the null string are
Is that right?
More information about the Ietf-languages