On 2-5-2010 3:59, Philipp Gesang wrote:
1. In sort-lan.lua, line 101 should read «['r'] = "r"», and line 144 «['r'] = 26, -- r».
i patched the file
2. Although I read the disclaimer about said file being “preliminary and incomplete” -- is there some rationale behind the range of integers for each language mapping? The mapping for English goes from 1 to 51, interleaving 2 integers for each letter (which is odd because it should start from index 3 with “a”, shouldn't it?), while the Czech one goes from 1 to 40 without skipping, Finnish and Austrian from 1 to 58.
some old (ruby) code was used etc etc
What about mapping them onto a larger but common scale that would alleviate multilingual sorting so that the alphabetical representation of the phoneme /a/ maps to the same value over different languages?† E.g. ["a"] = 3, -- in a Latin mapping, ["α"] = 3, -- in Greek mapping, ["а"] = 3, -- in a Russian mapping.
hm, interesting ... feel free to reshuffle and provide patches
† I know this is impractical for many writing systems and even within the set of Latin or Greek based alphabets it largely depends on a given purpose how much precision you need in sorting.
indeed but we can have multiple variants and are not bound to specific conventions Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------