[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Debian doesn't have to be slower than time.

#include <hallo.h>
Marco d'Itri wrote on Mon Feb 18, 2002 um 01:40:45PM:

> Please do not confuse i18n and l10n with unicode support.
> The reality is that most users do not need unicode support and cannot
> use it as their native encoding because doing this would fuck up
> communications with other people still using a local charset as their
> native encoding.

a) Most users use more-or-less up-to-date versions of MicroShit Outlook
(Express), so they can read unicode.

b) Good MUAs like mutt can be configured to try to recode the message
into a native charset and only use unicode if no NLS charset was

> And forcing it to be a release goal would be just stupid, unless you
> plan to do the all the work needed to make all programs support it
> correctly without breaking other things.

This is a good goal, but not easy to be realized. Though we should look
how to prepare Debian for this, since other OS are already going this
way. I suggest following steps:

In Woody+1: 
 - All not-localised programms must not break on UTF8
 - Allready localised programs must handle from/to unicode conversion
   transparently. This does not cover correct internal data processing,
   just user interaction.

In Woody+2:
 - Programs must process data correctly (almost working example: mutt)
   and provide transient solutions
 - Programs must switch the settings relative to used charset
 - using UTF8 as the default charset

 - is it possible to use a normal Linux console (without framebuffer or
   other crapy slow solutions) and a customised subset of Unicode as
   character map? I mean: I need only 128bit ASCII, few chars from
   iso-8859-15 and about 60 chars from koi8 charset. All this should fit
   in 255 chars.

"Und deshalb sollte man ja auch immer _alle_ Editoren installieren:  XEmacs
_und_ Vim."                                                  -- Sven Guckes

Reply to: