Hello,
somebody has asked for a survey of national sorting rules.
(I have already deleted that note.)
Now it has occurred to me that the IBM National Language Technical
Centre has produced a "National Language Information and Design Guide"
comprising such a list (in Volume 2, Chapter 3). I have only the
1st edition of July 1987, which still contained many inaccuracies.
I do not know whether these books have been overhauled, or whether
they still are available. Check with your local IBM representative;
order number was SE09-8002-00 (back in 1987).
Am 2000-06-15 um 08:54 h (GMT-0800) hat Jarkko Hietaniemi geschrieben:
> I think somebody just mentioned that many Italians like "i" and "j" to be
> "equal".
German DIN 5007 (of April 1991) also has its peculiarities.
It defines two, mutually incompatible, orders, one "for lists of proper
names" and the other one for anything else. The former one is used by
the descendants of the former German Postal Authority for their directories
(telephone, zip-codes, &c.), and I think, this is the reason that it
is still in the national standard (though with a note that it may be
dropped in later versions of the standard); the latter one (or variants
thereof, as exemplified below) is used in encyclopaedias and dictionaries.
In 1st approximation, the former equates "�", "�" und "�" to "ae", "oe",
and "ue", respectively, the latter to "a", "o" and "u". Bothe equate "�"
to "ss", and "�" to "th", in 1st approximation. (The "�" ruke is an
interesting quirk, as this letter is alien to German orthography.)
Both are three-level sorts: 1st-approximation homographs are sorted,
in 2nd approximation, by the letter variants, such as diacritical marks.
Here, Umlauts ("�", "�", "�") and the Eszet ("�") go after base characters
but before any other variant; DIN 5007 even defines a particular order
for accents and similar diacritical marks (based on their glyphical
appearance).
2nd-level homographs are sorted according to case: here, capitals go
after small letters.
Encyclopaedias and dictionaries usually comply with the 1st level. Some
state explicit, deviating rules for the 2nd, or 3rd, level (e. g. Duden
has "�" before "ss", in the 2nd level); some do not state their rules
in that detail, but nevertheless deviate in 2nd, or 3rd, level. Some
dictionaries sort 1st-level homographs on part of speech.
Best wishes,
Otto Stolz
