Summary: de-DE-1996 is better than de-1996-DE

[J.Wilkes]

> Dear language specialists,
> 
> Rather than continuing too much on the current threads, I have
> tried to sum up the arguments for my (and others') position
> that de-DE-1996 is better than de-1996-DE (not necessary in a
> particular order):
> 
> - In various fields of discourse (e.g. food, legal), the differences
>    between DE/AT/CH break mutual intellegibility. There is no such thing
>    for the difference between 1996 and 1901.
> 
> - The difference between DE/AT/CH is relevant in speech and writing
>    (don't confuse this difference with dialects); the difference
>    between 1996 and 1901 is only relevant in writing.
>
> - People (and in some cases systems, see Torsten's Mozilla example)
>    are already used to see the country identifier in second place.
>    Indeed, the specification says that a two-letter code in second
>    place is a country code. It is knowledge that we can and should
>    use.

I agree with these points in general.
About software implementations and RFC, I just posted something.

> - At least between DE/AT on the one hand and CH on the other hand,
>    there are clear orthographic differences (the absence of the sharp s
>    in the CH variant) in both 1901 and 1996 orthographies.
>    [And no, the sharp s hasn't been abolished in DE-1996 and AT-1996,
>    only reduced in usage.] So even a model that puts orthography before
>    vocabulary doesn't give a clear preference for putting the years first.

If the order was to mean something, I would not agree.
If the part that changes less often goes before the part that changes more often/has 
less value, as we know it from date and time or currency,
orthography would have to come first, and vocabulary after it.

Since RFC 3066 is flawed IMHO, consistency and logic are not paramount in 
defining these tags, unfortunately.
 

> - HTTP content negotiation. This is designed so that it is easy to
>    match a request for de-CH with a document labeled as de-CH-foo,
>    but not easy to match a request for de-CH with a document
>    labeled de-foo-CH. While I guess that both content negotiation
>    on DE/AT/CH and on 1996/1901 will be relatively rare, I think the
>    case for negotiation on DE/AT/CH is significantly stronger than the
>    case for negotiation on 1996/1901.

I don't agree, but rather leave this aside. I don't accept it as separate reason, but 
rather a subordinate case in the standard/implementation perspective.

> - At this time of change, the difference between the orthographies
>    receives considerable attention. In a few years, as a few years
>    ago, this difference will be mostly forgotten. The country-specific
>    differences won't disappear very soon, and won't get forgotten.

I don't think so, Unless you consider 95 years "a few".
Also, "Prediction is difficult--especially of the future".
 
> - There has been a claim that all authors write with the intent
>    of either using 1901 or 1996 orthography. That's probably true
>    for people publishing and for schoolchildren. Otherwise, people
>    don't consiously write to a specific orthography the same way
>    they don't write conciously to a specific vocabulary.

It also holds true for book authors, jurists, biblophiles, and especially for poets and 
lyricists!

The works of Benn, Rilke, Fried, Jandl, Arp, Arno Schmidt and many more will 
simply not work anymore with the 1996 orthography. When read in a 1996 
orthography scope, they get skewed and distorted; and their beauty vandalised.

The average person maybe does not choose consciously which orthography to use, 
applying a sketchy mixture instead.
Except for at least 230,000 persons in Schleswig-Holstein and an undetermined 
number for the other parts of germany.

The claim that "all authors" chose a specific orthography may not be true, but it is 
neither a small minority which does, nor are they unimportant or less influential. On 
the contrary, those whose words matter most, chose the orthography consciously.

While I agree with you in regard to adopting a certain order of tags, I oppose this 
specific reason.  


Johannes Wilkes
-- 
metabit * software and networks * heterogenous,distributed,generative  
Fon:(+49)228/242488-0 * Fax: (+49)228/242488-7
address: Kurfürsten-11 * D-53115 Bonn * Germany