For dialects that do not satisfy criteria we might establish for adding a
separate language code, we need to consider a mechanism for extension.
This came up in regard to the Norwegian codes (although not dialects I
realize, but I felt that a similar solution was needed). In March I
circulated the proposal below. Some reacted that we had already approved
the Norwegian codes and couldn't change them and that this was a major
change to the standard.  I do not think that this can wait until the
standard is up for review in 3 years, since we will certainly have to deal
with dialects in the meanwhile.

I propose that we consider this mechanism, and if we can agree upon it,
that we pursue establishing an amendment to the standard. This would
involve balloting the change with the member bodies.

The examples below are for Norwegian, but you could substitute any dialect
of another language (e.g. Valencian as a dialect of Catalan):

cat-val (ISO 639-2)
ca-val (ISO 639-1)

Please comment.


---------- Forwarded message ----------
Date: Thu, 2 Mar 2000 16:23:28 -0500 (EST)
From: Rebecca S. Guenther <[log in to unmask]>
To: [log in to unmask]

To ISO 639/JAC:

We had a meeting here at the Library of Congress of catalogers who are
applying the language codes, and the difficulty of applying the three
Norwegian codes was made clear.  It was felt that at least in the U.S.
and probably most places outside of Norway those applying the codes may
not have the expertise to be able to determine and it may not be
desirable for searching and retrieval purposes to make such fine
distinctions. There is also the problem that there exists two forms of the
language, one based on Danish and one on Old Norse (this taken from
Havard's ISO/DIS 639-l Annex C), while there are now 3 codes. As Havard
stated at our meeting, this is a national rather than international need,
although all of these codes are needed in Norway.

This is a case that is bound to come up time and time again in this
standard, when local needs conflict with international needs.  Those of us
maintaining large bibliographic databases and producing large numbers of
records may not be able to or want to make the fine distinctions that
might be made in the countries where the language is spoken.

Therefore I would like to propose the following solution.  The new codes
that were approved (nno and nob) would be appended onto the more general
code for Norwegian.  Thus, a hierarchical type of coding would be used:
        nor-nno  Norwegian Nynorsk
        nor-nob  Norwegian Bokmaal
For the alpha-2 list we would do the same, although I would argue that
only the alpha-3 code would be needed as an extension:
        no-nno  Norwegian Nynorsk
        no-nob  Norwegian Bokmaal
An alternative could be using the alpha-2 code as the second part:

We could consider applying this mechanism in the future where needed for
these types of situations, but we would NOT go back and look at the codes
we have as to whether others are similar.

This solution would be consistent with the ISO 3166 subentity codes, where
the code for a subentity is attached to a country code to be more
specific.  It would also be consistent with the direction in the current
proposed revision to RFC1766, where it is stated that a subtag may be used
in conjunction with a language code (example is: no-nynorsk, no-bokmaal).
The only difference is that a code would be used for the subtag, rather
than a language name.

I would like to open up discussion on this proposal during the next
several days. Please consider it and comment between now and next Tuesday,
7 March. Then I will send out a voting form and we will vote on it.

Although we have previously voted on these codes, I don't see this
possible change as a complete reversal of that decision, but a
reformulating of how the codes are presented.

I look forward to your comments.