Re: [NTG-context] sort order in registers (indexes)

Am 2018-06-10 um 14:11 schrieb Henning Hraban Ramm :

Am 2018-06-10 um 13:07 schrieb Henning Hraban Ramm :

...

...

I think this may achieve what you want: \setupregister[index][method={zm,zc}] I hope it helps,

Thank you very much!

I overlooked that there is indeed some documentation: http://wiki.contextgarden.net/Command/keyword:method

But reading through that and the source I’m still confused.

I’d like to write a proper explanation for the wiki (and my book).

Is there anywhere documentation about the meaning/goal of the presets or algorithms?

Are these codes translatable into something like "ignore diacritics", "ignore upper/lowercase" etc.?

Ok, I think I got it... For a proper sorting, you first apply a "mapping", then a "casing" and finally "unicode". Presets: default = upper like lowercase, diacritics separate before = upper before lower, diacritics ignored after = lower before upper, diacritics ignored first = lower before upper, diacritics separate last = upper before lower, diacritics separate * If you don’t set the sorting method, the preset "first" ist used (and not "default"). * There’s no preset for the (in my eyes most meaningful) combination "upper like lowercase, diacritics ignored" (zm,zc,uc). * Aren’t language specific sorting rules possible at the current state? Or does "unicode" handle that? E.g. -- DIN 5007-1 (German default sorting) is like zm,zc,uc, but ß should be sorted like ss. -- DIN 5007-2 (German phonebook sorting) would additionally require umlauts to be sorted as ä = ae etc. -- Austrian phonebook sorting sorts umlauts after base vocals, i.e. a, ä, o, ö, u, ü, s, ß. -- Danish and Norwegian: x, y, z, æ, ø, å -- Finnish and Swedish: x, y = ü, z, æ, ä, ö, ø, å (until 2006 v = w) -- etc. (according to https://de.wikipedia.org/wiki/Alphabetische_Sortierung) If nobody objects I’ll add this to the wiki. Greetlings, Hraban --- https://www.fiee.net http://wiki.contextgarden.net https://www.dreiviertelhaus.de GPG Key ID 1C9B22FD