W3C home > Mailing lists > Public > www-international@w3.org > April to June 2008

Re: 2 many language tags for Norwegian

From: Frank Ellermann <nobody@xyzzy.claranet.de>
Date: Wed, 30 Apr 2008 15:26:40 +0200
To: www-international@w3.org
Message-ID: <fv9rtf$rno$1@ger.gmane.org>

Leif Halvard Silli wrote:  

> Let me rephrase the above: German speakers can be
> lucky that, on the Web, their different flavours
> of German, are tagged using geographical subtags.

German speakers living in the region where frr is
used will have to use frr +/- nds +/- da + de to
get what they want (at some point in time, the last
time I looked at frr sites they did not yet use the
relatively new frr code)

Ditto for all other official or simply popular (tr)
languages used in DE.

> And they can all be tagged as de-something.

No, you can't tag frr or nds as de-something.  You
also won't try workarounds like fy-DE for frs, or
fy-DK for frr as used in DE, or nl-DE for nds, not
after the better tags were registered (fy-DE and
nl-DE likely exist, and do NOT mean frs or nds).

> That standards is doing things opposite of the
> way Apache works.

I've not the faintest idea how language negotiation
works with Apache, but if it cannot deal with more
than one Accept-language it is broken.  The version
mentioned by you (1.3) is rather old, isn't it ?

> So are the standards developed in the free?

You are free to subscribe to the LTRU list, and to
participate, yes.  One of the folks replying to your
posts is an active participant of that WG, another
is a co-Chair, and I'm a former participant.  They
will be delighted when you reopen the extlang issue
(not).

> [old] sgn-DK ([new] Preferred-Value: dsl)

Good idea, sgn-DK etc. was always shaky as there can
be more than one sign language per region (= country).

> I suppose that sign languages have similarities.

AFAIK no.  They might have the same signs for letters
or phrases, but when you're down to letters it is in
the relevant language like da.  You could also say
that da, fo, is, nb, nn, no, and sv are "similar" and
should be grouped under a Scandinavian "macrolanguage".

> And the way I imagine progress, one should be able
> use that in language tags as 'nor-nbo'.

No progress, the "nor-" bit in nor-nbo is redundant,
it makes no sense to put nor-nbo in billions of pages.

>> 26*25 is not good enough for all languages worldwide.
> Not sure what 26*25 refers to.

alpha2:  26 * 25 letters, I confused this with country
codes, XA..XZ are private use.  I should have written
26 * 26, sorry.

> gsw. Swiss German. Added in 2006, according to the
> language subtag registry.

> So why no use de-CH instead?

Because de-CH is "high German" as used in the NZZ with
some Swiss variations (no eszet, and some words such as
"velo" or "zensurieren" not used elsewhere), while gsw
is what they use when they talk.  BTW, gsw is also used
outside of CH in FR and DE.

>> Let alone "frr", "frs", "dsb", "lsb", etc. 
> Those are not the German language.

Minority languages used mostly in Germany.  Like se for
Norway (don't beat me if that is not the relevant Sami
language for Norway, I know nothing about Sami).  For
speakers of these languages they will add more common
languages (de+hsb for say dsb, no+nb+nn for se-NO) to
their browser configurations.  (s/lsb/hsb/, my bad).

> (They are German/Germanic languages.)

dsb+hsb are no Germanic languages, they are languages of
Germany (the Anglosaxon name for "Germany" is weird, of
course Norway is more Germanic than most other countries
of the world, but maybe FO could beat NO in this respect)

> (For the record, I have not ever proposed that 'no-SE'
> should be valid.)

ACK, you said you found it on a government site, right ?

> There are 3 Frisian languages. I would say the same:
> They should have a common mother tag for "plain" Frisian.

Unfortunately plain Frisian was lost hundreds of years ago.
Like en is not more ang, and speakers of nds won't grok
ang, they also won't grok en unless they learned it.  

It's unnecessary to find a Saxon macrolanguage for ang, en,
nds, sxu, etc., or a Frisian macrolanguage, that's all old
history, like a Scandinavian macrolanguage => forget it, it
won't help you to avoid to patch your Apache, or to avoid to
find the place where Safari hides its configuration for more
than one Accept-Language.  I'd bet that it can do, you just
havent't found it yet :-)  Netscape 3 vintage '96 supported
more than one Accept-Language.

> I think 'de-nds' would have been smarter than only 'nds'.

The original idea in RFC 1766 was a variation of "locales", 
language + region.  And the Norwegian author of RFCs 1766
+ 3066 knew everything about the fine print of "no".  The
concept was later extended, RFC 4646 is new, ISO 639-3 is
very new, 4646bis is a draft, for ISO 639-6 I lost track...

 Frank
Received on Wednesday, 30 April 2008 13:24:50 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 19:17:17 GMT