Subject: Re: national language support
To: None <email@example.com>
From: Andreas Gustafsson <firstname.lastname@example.org>
Date: 10/24/1994 11:08:50
> I just wondered, if national language support will be added to
> The basic support for it is being worked on, but except for the ctype
> changes (e.g. calling setlocale()), no programs have been modified to
> use it. This involves several things; primarily converting programs
> to actually *use* the locale information for, e.g., the decimal point
> character, and converting most everything to use message catalogs.
> Also, someone has to actually translate the message catalogs into other
The original poster also mentioned 8-bit cleanliness of user commands.
That issue is mostly orthogonal to that of locales, and I personally
think it would be the more useful one to spend some effort on.
As a native Swedish speaker, I have to struggle daily with utilities
like "more" which are not 8-bit clean (meaning that they are unable to
deal with three of the letters in the alphabet), and sometimes even
with programs that unconditionally reset my terminal to 7 bits / even
On the other hand, my experience with locales is that they typically
go unused because using them would cause too many programs to break.
For example, text-based file formats become unportable as decimal
points unexpectedly turn into commas. To me, any effort spent on
locales is wasted (although it may still be judged necessary just to
be able to claim standards conformance).