[gentoo-dev] Re: enable UTF8 per default?

[Mathieu Bonnet <mathieu.bonnet@nalkym.org>]

Hi,


Patrick Lauer <patrick <at> gentoo.org> writes:
>
> Enabling the unicode useflag in the profiles should help our
> international users and should not cause any problems. [...]
> 
> On a tangent I wonder if pulling in extra fonts as a dependency of X
> makes sense (useflag controlled, enabled by default) - that way the
> unicode capabilities are available without any configuration.
>


Don't forget about the kernel configuration, the locales configuration, and
individual apps configuration (run-time configuration, I mean).

See: http://www.gentoo.org/doc/en/utf-8.xml

Having all this done automatically would indeed be nice.


>
> Are there any known bugs / problems this would trigger? Any
> reasons against that?
>


The classic problem is that non-ASCII characters will not be readable by the
people still not using UTF-8, instead of their old ISO/other charset...

If someone is using an old ISO/other charset and someone else is using UTF-8,
and if they speak the same language, they won't be able to understand
eachother's non-ASCII characters... (and most users won't understand why -and
most users will blame it on the configuration of the user using UTF-8 :/)

Classic case of incompatibility hell...

UTF-8 is here to solve this, but it only works if everybody changes to UTF-8...
but as long as not all software support UTF-8 (sometimes, they don't, only
because the devs don't feel like it is necessary :/), and as long as not every
users use UTF-8, then it will be perfectly respectable to stick to ISO/other
charsets, to avoid problems with family members, friends and colleagues who
still use ISO/other charsets and don't want to change or just can't change...
(Windows users, mostly). (This is pretty similar to the situation concerning
instant messaging... you want to go to Jabber, but everyone else is using MSN,
AIM, ICQ and Yahoo...).

This is why we should still not go with fully activating UTF-8 as default...
what we should do, is ease the migration to UTF-8, for everyone who feel like
using it... (and for everyone whose family members, friends and colleagues
already use UTF-8 -but it should be quite rare). If it's easy, more and more
people will migrate, and when Windows will activate UTF-8 by default (let's pray
MS will feel like implementing this correctly), and most Windows users have
updated (or went to Linux/BSD/other ^_^), then we should switch to all UTF-8 by
default.

Until then, as said, the migration should be easier for the user... it means we
should also take care of run-time configuration, maybe with a new USE flag, to
preconfigure every programs which support UTF-8, to use it by default (instead
of just supporting UTF-8, with the "unicode" USE-flag). Of course, the kernel
and locales should also be configured automatically.

Then, another great thing would be to be able to easily press the backpedal and
goes back to ISO/old charset, if too many incompatibility problems arise... (if
it's hard to come back, and they have problems, people will hate UTF-8 for years
:/ -and they'll tell other).

Instead of an USE-flag for default preconfiguration, maybe a script could take
care of modifying the user programs configuration, to easily activate and
deactivate UTF-8 by default, in every supported programs... (well, the script
would also allow to configure programs one at a time).

We should also grealty document the migration process and the incompatibility
problems which may arise if you try to communicate with people using ISO/old
charsets... (and grealty explain them the use of `iconv` and similar programs).

Another thing is to list every packages which do not support UTF-8... (editors,
like NEdit, for example), and maybe make `emerge` warn the user, if he uses
UTF-8, that the package he is trying to emerge, do not support UTF-8... (well,
an ewarn, at the end of the emerging process would be enough, but to be warned
before emerging the package would be nice).

Anyway, we should separate "activating support for Unicode/UTF-8" and
"activating use of UTF-8, by default, in every programs", and we should support
both...

If they can be cleanly be separated, we should probably activate support for
Unicode/UTF-8 by default, now... but as said, activating use of UTF-8 by default
for all program should not be done until quite some time...


Cya.


-- 
gentoo-dev@gentoo.org mailing list