Hi again.
Im going also to implement a translations for EPIC,
and want to ask USERS for suggestions and ideas.
As for me, for russian language,
it will be enough to have charset parameter for each window,
default window charset and terminal charset.
Setting proper charset according to specific servers or channels
is probably a script task.
I want to make non 1:1 mapping and with 'fallbacks',
which allows to render characters missed in terminal charset,
like egg. links or GNU recode do.
However, there are other problems in other languages.
- Entering digraphs. I never use them and so, have no clue
how is it better to implement this.
i've seen two ways:
having special combine key, and autocomposing
like autosubstitution ":o" with o umlaut,
and ": " (':'+space) with single ":"
Maybe to suport both ? or leave it for scripts ?
- Non-8bit encodings.
Japanese, Chinese and similar languages use either 2-byte encoding,
or some scripts to encode/decode text into ascii.
There exist several 2byte 'charsets' as well as several scripts, IIRC.
As for Japanese - i've seen different channels using different methods
even under the same OS.
Here i'm completely giving up, having never seen how such languages
are used in console (are they used at all? :)
My only suggestion is to make parameter 'encoding' in addition
or better instead of 'charset' and support something like
'plugins' to translate "user-defined" encoding types,
one for 8bit charsets should be built in.
This require some API for translation module.
Or - maybe leave all these for iconv and do not panic ?
But [ig]conv, afaik, does not support scripting encodings,
as well as transliteration (aka romanji for Japanese).
qMax
_______________________________________________
List mailing list
[EMAIL PROTECTED]
http://epicsol.org/mailman/listinfo/list
