05 Mar 17 20:12, you wrote to Michiel van der Vlist:
KK> Sadly, that method wouldn't work for cyrillic names as there are several KK> standards for names transliteration. Wiki mentions that there are at least KK> five. Above that users often prefer using mixed versions and add something KK> from themselves like 'v' ending becomes 'off'.
I have been looking into Romanisation of Cyrillic names. Indeed there are a number of standards. Just to make a start, I looked at the ICAO tranlation, that is used to translate Cyrillic names to Latin for use in machine readable passports.
The is some Perl code to do the translation from Cyrillic to Latin. I tried a reverse routine, that is not complete yet, ad it probably does a dusgusting translation, but at lest the same comes out when translated back to Latin. A norn for Fidonet could be to use the ICAO method for the nodelist as well.
KK> Yes. So maybe for now the UTF-8 version needs to be derivative version of KK> ascii version, not the completely parallel replacement. It must be KK> generated from ascii segments and additional information like native KK> names, locations and so on.
One way or the other.
KK> I hardly imagine a man who in 2017 doesn't know latin alphabet on KK> transliteration level. But easily imagine the one who doesn't know KK> cyrillic alphabet. I don't even say anything about any asian languages or KK> for example hebrew.
Lets take it one step at a time. There are very few participants in Fidonet that require Asian languages. Cyrillic is difficult enough with the difference between Russia, Ukrain, en Belarusse.
As for hebrew, all sysops in R40 are of Russian origin.
KK> Yes, so as I suggested above we can use ASCII version as a source for KK> "extended" unicode version. And virtually nothing prevents generation of KK> additional versions with different format.
It is far easier to degrade an extended version, that it is to enrich a limited one.
Kees
--- GoldED+/LNX 1.1.5 * Origin: As for me, all I know is that, I know nothing. (2:280/5003.4)
