Notes on the Free Translation Project
Free software is going international! The Free Translation Project
is a way to get maintainers of free software, translators, and users all
together, so that will gradually become able to speak many languages.
A few packages already provide translations for their messages.
If you found this `ABOUT-NLS' file inside a distribution, you may
assume that the distributed package does use GNU `gettext' internally,
itself available at your nearest GNU archive site. But you do *not*
need to install GNU `gettext' prior to configuring, installing or using
this package with messages translated.
Installers will find here some useful hints. These notes also
explain how users should proceed for getting the programs to use the
available translations. They tell how people wanting to contribute and
work at translations should contact the appropriate team.
When reporting bugs in the `intl/' directory or bugs which may be
related to internationalization, you should tell about the version of
`gettext' which is used. The information can be found in the
`intl/VERSION' file, in internationalized packages.
One advise in advance
If you want to exploit the full power of internationalization, you
should configure it using
to force usage of internationalizing routines provided within this
package, despite the existence of internationalizing capabilities in the
operating system where this package is being installed. So far, only
the `gettext' implementation in the GNU C library version 2 provides as
many features (such as locale alias or message inheritance) as the
implementation here. It is also not possible to offer this additional
functionality on top of a `catgets' implementation. Future versions of
GNU `gettext' will very likely convey even more functionality. So it
might be a good idea to change to GNU `gettext' as soon as possible.
So you need not provide this option if you are using GNU libc 2 or
you have installed a recent copy of the GNU gettext package with the
Some packages are "localizable" when properly installed; the
programs they contain can be made to speak your own native language.
Most such packages use GNU `gettext'. Other packages have their own
ways to internationalization, predating GNU `gettext'.
By default, this package will be installed to allow translation of
messages. It will automatically detect whether the system provides
usable `catgets' (if using this is selected by the installer) or
`gettext' functions. If neither is available, the GNU `gettext' own
library will be used. This library is wholly contained within this
package, usually in the `intl/' subdirectory, so prior installation of
the GNU `gettext' package is *not* required. Installers may use
special options at configuration time for changing the default
behaviour. The commands:
will respectively bypass any pre-existing `catgets' or `gettext' to use
the internationalizing routines provided within this package, enable
the use of the `catgets' functions (if found on the locale system), or
else, *totally* disable translation of messages.
When you already have GNU `gettext' installed on your system and run
configure without an option for your new package, `configure' will
probably detect the previously built and installed `libintl.a' file and
will decide to use this. This might be not what is desirable. You
should use the more recent version of the GNU `gettext' library. I.e.
if the file `intl/VERSION' shows that the library which comes with this
package is more recent, you should use
to prevent auto-detection.
By default the configuration process will not test for the `catgets'
function and therefore they will not be used. The reasons are already
given above: the emulation on top of `catgets' cannot provide all the
extensions provided by the GNU `gettext' library. If you nevertheless
want to use the `catgets' functions use
to enable the test for `catgets' (this causes no harm if `catgets' is
not available on your system). If you really select this option we
would like to hear about the reasons because we cannot think of any
good one ourself.
Internationalized packages have usually many `po/LL.po' files, where
LL gives an ISO 639 two-letter code identifying the language. Unless
translations have been forbidden at `configure' time by using the
`--disable-nls' switch, all available translations are installed
together with the package. However, the environment variable `LINGUAS'
may be set, prior to configuration, to limit the installed set.
`LINGUAS' should then contain a space separated list of two-letter
codes, stating which languages are allowed.
Using This Package
As a user, if your language has been installed for this package, you
only have to set the `LANG' environment variable to the appropriate
ISO 639 `LL' two-letter code prior to using the programs in the
package. For example, let's suppose that you speak German. At the
shell prompt, merely execute `setenv LANG de' (in `csh'),
`export LANG; LANG=de' (in `sh') or `export LANG=de' (in `bash'). This
can be done from your `.login' or `.profile' file, once and for all.
An operating system might already offer message localization for
many of its programs, while other programs have been installed locally
with the full capabilities of GNU `gettext'. Just using `gettext'
extended syntax for `LANG' would break proper localization of already
available operating system programs. In this case, users should set
both `LANGUAGE' and `LANG' variables in their environment, as programs
using GNU `gettext' give preference to `LANGUAGE'. For example, some
Swedish users would rather read translations in German than English for
when Swedish is not available. This is easily accomplished by setting
`LANGUAGE' to `sv:de' while leaving `LANG' to `sv'.
For the Free Translation Project to be a success, we need interested
people who like their own language and write it well, and who are also
able to synergize with other translators speaking the same language.
Each translation team has its own mailing list, courtesy of Linux
International. You may reach your translation team at the address
`LL@li.org', replacing LL by the two-letter ISO 639 code for your
language. Language codes are *not* the same as the country codes given
in ISO 3166. The following translation teams exist, as of August 1997:
Chinese `zh', Czech `cs', Danish `da', Dutch `nl', English `en',
Esperanto `eo', Finnish `fi', French `fr', German `de', Hungarian
`hu', Irish `ga', Italian `it', Indonesian `id', Japanese `ja',
Korean `ko', Latin `la', Norwegian `no', Persian `fa', Polish
`pl', Portuguese `pt', Russian `ru', Slovenian `sl', Spanish `es',
Swedish `sv', and Turkish `tr'.
For example, you may reach the Chinese translation team by writing to
If you'd like to volunteer to *work* at translating messages, you
should become a member of the translating team for your own language.
The subscribing address is *not* the same as the list itself, it has
`-request' appended. For example, speakers of Swedish can send a
message to `firstname.lastname@example.org', having this message body:
Keep in mind that team members are expected to participate
*actively* in translations, or at solving translational difficulties,
rather than merely lurking around. If your team does not exist yet and
you want to start one, or if you are unsure about what to do or how to
get started, please write to `email@example.com' to reach the
coordinator for all translator teams.
The English team is special. It works at improving and uniformizing
the terminology in use. Proven linguistic skill are praised more than
programming skill, here.
Languages are not equally supported in all packages. The following
matrix shows the current state of internationalization, as of August
1997. The matrix shows, in regard of each package, for which languages
PO files have been submitted to translation coordination.
Ready PO files cs da de en es fi fr it ja ko nl no pl pt sl sv
bash |    | 3
bison |    | 3
clisp |     | 4
cpio |      | 5
diffutils |      | 5
enscript |       | 6
fileutils |           | 10
findutils |         | 8
flex |     | 4
gcal |      | 5
gettext |           | 11
grep |          | 9
hello |           | 10
id-utils |    | 3
indent |     | 4
libc |        | 7
m4 |      | 5
make |       | 6
music |   | 2
ptx |         | 8
recode |          | 9
sh-utils |        | 7
sharutils |      | 5
tar |           | 10
texinfo |  | 1
textutils |          | 9
wdiff |         | 8
16 languages cs da de en es fi fr it ja ko nl no pl pt sl sv
27 packages 3 2 24 1 17 1 26 2 1 11 20 9 19 7 7 17 167
Some counters in the preceding matrix are higher than the number of
visible blocks let us expect. This is because a few extra PO files are
used for implementing regional variants of languages, or language
For a PO file in the matrix above to be effective, the package to
which it applies should also have been internationalized and
distributed as such by its maintainer. There might be an observable
lag between the mere existence a PO file and its wide availability in a
If August 1997 seems to be old, you may fetch a more recent copy of
this `ABOUT-NLS' file on most GNU archive sites.