bitlord
commented
9 years ago Each table could have header with some basic info. Example
lang=sr info=transliterate all cyrilic serbian letters to latin serbian letters a a B B ...
And the trliterator would be able to auto-select table based on .po file that is parsed
Autodetection would be nice, I was thinking about few solutions for table files and how to handle selection, autodetection is probably the most elegant way of doing it for the user. For table files I would like to leave them as clean as possible (not sure what can come later with other language support. Also table metadata can be useful, like info= you provided), so adding extra data in it is maybe not needed, but just using the file naming scheme, for which I still don't have right solution.
For storing tables I think tables/ is ok, with every table in separate file with possible file naming scheme (including standardized language code in the name (sr_
For selecting languages
- Autodetection (based on .po files "metadata" I guess) (This is probably the best option, or should be default (if possible) with other/s described as fallback)
- cmdline opts --inscr <lang/script_code> --outscr <lang/script_code>
- for people doing localization/translation they mostly use one language (config file ~/.trltr or ~/.config/.trltr)
- using 'table' as a default (currently used) but looking for it in './tables/table' where 'table' is a symlink to the table you want to use
This is not final, just ideas, every comment/suggestion is welcome!
btw. I'm not very good with the theory how things work with languages, translation, ... if I don't understand the topic, please try to point me to good documentation/article where I can read and learn about it ;-)
Each table could have header with some basic info. Example
And the trliterator would be able to auto-select table based on .po file that is parsed