Convert to UTF-8

[SourceForge/phpwiki.git] / locale / README

This document describes how to add a new language translation to PhpWiki.

If you just want to use one of the existing language translations, see
part Four in 'config/config.ini' for the DEFAULT_LANGUAGE or your special
starter script (wiki or index.php) to change the current or default language
of your Wiki. Modify or add a line that sets $LANG to use the two-letter
code of one of the supported languages, like this:

    // Select your language/locale - default language "en": English
    // English "en"  (English    - HomePage)
    // Dutch   "nl" (Nederlands - ThuisPagina)
    // Spanish "es" (Español    - PáginaPrincipal)
    // French  "fr" (Français   - PageAccueil))
    // German  "de" (Deutsch    - StartSeite)
    // Swedish "sv" (Svenska    - Framsida)
    // Italian "it" (Italiano   - PaginaPrincipale)
    // Japanese "ja" (Japanese   - ホームページ)
    // Chinese  "zh" (Chinese   - 首頁)
    $LANG="it";

Phpwiki uses the DEFAULT_LANGUAGE internally, when no $GLOBALS['LANG'] is given.
With certain starter script which set $LANG you can easily provide a multi-lingual wiki.
For convenience provide some InterWikiMap entries for easier switching to the other 
languages then. See http://phpwiki.org/MultiLingualWiki

Phpwiki uses GNU gettext tools to provide and maintain multi-lingual
messages for different languages. Even if you are already familiar
with xgettext you will want to read this document to be aware of
translation issues and tips specific to PhpWiki; perhaps skimming
through the Makefile section.

PhpWiki does not require gettext support to be compiled in to PHP. It
automatically provides a pure PHP replacement in case it's not
available.


Formalities
===========
The modern English language has no clear rules for differentiating
between the formal and informal use of the spoken word. This both
simplifies and complicates matters for translators, as many languages
do still make such a distinction. For the most part, PhpWiki is written
using the casual forms of messages and explanatory text--after all the
WikiWikiWeb is an open and friendly place. :-)

However, there is no reason why translations of Wiki commands like
"Edit" or "FindPages" should not be written formally. For the sake of
recognition or clarity when the word function is the same as another
common computer term or menu-item, if you feel it would be more
effective to employ a formal variation then please do so. When the
formal word is significantly longer than the informal word, make a
judgement call or substitute an abbreviation, provided that it will
be easily recognised.

Remember that a good and thorough translation is a subjective and
collective process. Once people have had a chance to test-drive your
newly-translated PhpWiki there will always be suggestions for
reshaping and improving it.

Diversity and Change are part of WikiEssence. By all means don't
belabour the translation process, and have some fun!


General Steps
=============
The general steps to creating a new language module for PhpWiki:

1. Copy and rename the master template to create a new file for your
   language ('phpwiki.pot' => 'xx.po').

2. Translate all the strings in the new language file ('.po file').
   Empty strings will be kept in english.

3. Run 'make' to automatically create or update any additional files
   ('.po' => '.mo' => 'phpwiki.php' files).

4. Create translations of the content for the default pages ('pgsrc'
   files).


Example
=======
Let's assume for example that you would like to add an Islandic
translation to PhpWiki.


Text Strings
------------
Duplicate the file phpwiki.pot in the 'locale/po/' folder and name it
'is.po' ('is' is the code for Islandic). Next, translate all the
strings inside 'is.po' from English to Islandic. Unix Hint: Emacs has
a handy "po translation mode" for you. See README.coding in the doc folder.


Character Encoding
------------------

Currently most of the language files are saved using the ISO-8859-1
character encoding to preserve accented characters. Make sure the text
editor you use is capable of performing the appropriate Latin-1
translation. Strictly speaking, ISO-8859-1 is *different* than Windows
code page 1252 or MacRoman. Upon quick inspection one will notice that
many of the letters do occupy the same positions in each of their'
respective encoding tables, so it is easy to understand how people can
make this false assumption.

Limitation:
PhpWiki is not able to translate between different character sets on the 
fly (iconv). Certain languages, like Japanese or Chinese use special character 
sets (euc-jp, utf-8), which must be defined in part four in config/config.ini.
Therefore other languages in different character sets will not be correctly 
displayed.
PhpWiki is not yet properly utf-8 safe. That means that certain valid utf-8 multichar 
sequences with \0 in it, will break the traditional php string or pcre functions or 
that the length of the string will be miscalculated.
There's an experimental utf-8 patch on the sf.net patches site if you want to try that.
See http://phpwiki.org/Utf8Migration

Word Reordering
---------------
Different languages have different word ordering requirements. When a
key word such as a person's name needs to be added to the end of a
sentence in english, there is no guarantee that the same word will
appear at the end of a sentence once translated to another language.

PhpWiki has been designed with this in mind. (Standard printf
notations like %s and %1$s, %2$s will work with PhpWiki, even though
reordering of variable-substitutions is not actually part of PHP).
(The printf(3) manual page on your system may be of some help.)


Take the following English phrase as an example. This message would be
displayed in the browser when someone wants to compare two revisions
of a Wiki page. The corresponding entry for the German translation for
this phrase (from the file 'de.po') reads:
 
    #: ../lib/diff.php:251
    #, c-format
    msgid "Differences between %s and %s of %s."
    msgstr "Der Unterschiedsergebnis von %3$s, zwischen %1$s und %2$s"

In the English version, PhpWiki will substitute the phrases "version
1", "version 2" and the name of the page being compared. The
placeholder '%s' indicates where text will later be substituted, while
'%d' is used to indicate a number will be inserted.

Sentence structure of German is different than English, and this case
dictates that the "page name" must come first, then followed by
"Verson 1." and finally "Version 2."

To summarize, when the word ordering must differ, insert "1$", "2$"
etc. into the middle of '%s' and repeat for each instance of '%s' in
the phrase.  If you use this "$" notation, you must use it for each
of the the format specifications within the string.

Here are a couple more examples. They are fictional but they serve to
illustrate how to handle special cases, should they arise:


#msgid "Page version '%d' named '%s' is not available."
#msgstr "Xyqlmn vvvm » %2$s « mnqee » %1$d « Gvbbbnbhkkk eeek."

The %s and %d are reversed, so '2$' and '1$' have been used to
indicate the new ordering. The punctuation marks for this language
dictate extra spacing, so this has also been accounted for in the
translation (i.e. the quote marks are only for emphasis, they are not
considered part of the %s placeholder).


#msgid "Page named '%s' version '%d' written by author '%s' is not available."
#msgstr "Qppn wwmoooppp '%3$s' vvvm '%1$s' mnqee '%2$d' Gvbbbnbhkkk eeek."

Name and author are reversed. Even though the position of the number
('%d') isn't changed, still it must be changed to '%2$d' --- if you
use a '$' on one of the format specifiers you must use it on all of
them.  The punctuation for this particular language is the same as
english so it is unchanged.


While translating the text strings if you are uncertain about the
syntax, look at the '.po' files of the other languages as an
example. If you are stuck or simply can't make any sense of all this,
just ask one of the PhpWiki programmers on the mailing list to help
out.

The 'phpwiki.php' files do not need to be created or edited because
the Makefile will create and update these files automatically. See the
'Makefile' section below.


Default Pages
-------------
Most of the work will be in the translation of the default pgsrc
files. As a starting point you can copy the English 'pgsrc' directory:

    mkdir locale/is
    cp -rv pgsrc locale/is

For these 'pgsrc' files it will be sufficient to change the page names
to Islandic, and maybe translate the HomePage and give it an Islandic
name. Again, for anything you don't know, look at the 'nl' or 'de'
versions.

The best approach to translating the default page content is to do all
of your page editing in the web browser itself, then perform a page
dump to save the pages as MIME text files.  Some of the pages are
locked so you will have to log into PhpWiki as the administrator
before you can edit them.

<FIXME>
   Add instructions for editing MIME headers of files before moving
   files into '/locale/is/pgsrc'

 - keep modification date, page name and lock, remove author.
   Use action=PageDump - "Preview as developer format" for this.
   Example:

   Date: Sun, 27 Jan 2002 17:28:55 -0500
   Mime-Version: 1.0 (Produced by PhpWiki 1.3.2-jeffs-hacks)
   X-Rcs-Id: $Id$
   Content-Type: application/x-phpwiki;
     pagename=Neueste%C4nderungen;
     flags=PAGE_LOCKED;
     charset=iso-8859-1
   Content-Transfer-Encoding: binary

 - Make sure to rename files with accents in the page name converted
   to uppercase hex codes as shown in the pagename= line of the
   Content-Type header.

   (e.g. "GästeBuch" => "G%E4steBuch")

 - Translate body text and rename plugin pages to match changes
   specified in the '.po' file. Leave plugin code as english.
   Example:

   <?plugin RecentChanges days||=30 show_all||=0 show_minor||=0 ?>

</FIXME>


Makefile
-------- 
The Makefile calls 'xgettext' to automatically perform a number of
important translation tasks:

 * Scans through all the source code and retrieves the english
   strings used as text arguments, then collects and indexes them into
   the file 'phpwiki.pot'.

 * Merges any new differences of the collected English text strings
   with similar text strings found during any previous runs of 'make',
   stored inside each of the translated '.po' files.

 * Makes note of which English text strings have been added, reworded
   or removed. The translated strings in the '.po' files are then marked
   as "fuzzy" translations for all cases where the English text has been
   changed. This makes it easy for translators to spot which items need
   to be updated. (Emacs' po mode also uses this information).

 * The necessary '.mo' files and 'phpwiki.php' text files are
   synchronized and sorted according to the translated contents of the
   '.po' files, for each of the locale subdirectories.

When a new language module is added to PhpWiki, the 'Makefile' in the
'locale' folder also needs to be updated. The process:

 0. Change into the locale directory (if you're not already there.)

 1. If there isn't one already, Create a new .po (in the po/ subdirectory)
    for your translations:

      cp po/phpwiki.pot po/xx.po

    where 'xx' should be the two letter character code for the language
    you are translating to.

 2. If you've created a new .po file (or if there are new .php source files
    containing translatable strings,) update the Makefile:
 
      make depend

 3. To make sure the list of translatable strings in the .po is in sync
    with the actual php source code:
 
      make po

 4. Edit the translations in the .po file.  If you have emacs available,
    emacs' po-mode is very helpful when doing this.

 5. Translators may include comments in the po file. If using a text
    editor other than emacs (with po-mode), add any comments BEFORE an
    entry, and prefix each line with # like this:

# This entry is important for French punctuation rules which stipulate
# a space before each colon.
#: ../lib/display.php:54 ../lib/plugin/PageGroup.php:91
#: ../lib/plugin/PageGroup.php:156 ../lib/plugin/PageGroup.php:167
#: ../lib/plugin/PageGroup.php:183 ../lib/plugin/PageGroup.php:190
#: ../lib/plugin/PageHistory.php:216 ../lib/plugin/UnfoldSubpages.php:164
#: ../themes/default/templates/head.tmpl:46
#, c-format
msgid "%s: %s"
msgstr "%s : %s"

 6. To update the compiled translation files (phpwiki.mo and
    phpwiki.php):

      make

Make will then automatically generate and update all the necessary
files. If this step fails because you don't have the necessary
software installed, then send your '.po' files to one of the PhpWiki
developers who will run Makefile for you.


NOTE: "Fuzzy" translations will NOT be included in the final '.mo' and
'phpwiki.php' files. If you are not using the emacs editor, make sure
to remove the fuzzy marks from any *completed translations* before you
run the makefile, like this:

    #: ../lib/plugin/BackLinks.php:44
    #, fuzzy, c-format
    msgid "These pages link to %s:"
    msgstr "%d pages pointent vers %s."

should become:

    #: ../lib/plugin/BackLinks.php:44
    #, c-format
    msgid "These pages link to %s:"
    msgstr "%d pages pointent vers %s."

and...

    #: ../lib/Toolbar.php:160 ../templates/browse.html:54
    #, fuzzy
    msgid "Lock page"
    msgstr "Bloquear página"

should become:

    #: ../lib/Toolbar.php:160 ../templates/browse.html:54
    msgid "Lock page"
    msgstr "Bloquear página"


HTML Templates
--------------
The template files do not need to be translated. As of PhpWiki 1.3 all
the text strings in the html templates are cross-referenced with the
translations in the '.po' files.

*** Note: Updating html template translations from PhpWiki 1.2 to 1.3: ***

The translated version of the tips for TextFormattingRules must be
moved from the old html template 'editpage.html', and placed into the
'pgsrc' for the default page of TextFormattingRules. A plugin now
extracts this text and inserts it when editing a page in PhpWiki,
rather than it being embedded within the html template itself.

It is suggested this paragraph would go at the top of the page. It
must be in a section heading entitled "Summary" in order for the
editpage template to find it. Of course you will substitute the
translations for "TextFormattingRules" and "Summary", according to the
wording you used for these phrases when you translated the '.po' file.

Refer to the English "TextFormattingRules" and German (de)
"TextFormatierungsRegeln" pages to see working examples.


Finale
======
After you have finished translating, you will want to see the result
of your efforts. Change the LANG setting in 'config.ini' to the
two-letter code for your language.

Et voilà, Phpwiki should now speak Islandic!

If your translation was a success, you may also want to add a
translation of these instructions for translating PhpWiki ;-)

About gettext
-------------
To learn more about GNU gettext and '.po' files, you may find some
information at:

    <http://www.iro.umontreal.ca/contrib/po/HTML/>
    <http://www.gnu.org/directory/gettext>

Good luck,
Jan Nieuwenhuizen <janneke@gnu.org>
Arno Hollosi <ahollosi@mail.com>
Carsten Klapp <carstenklapp@users.sourceforge.net>
Reini Urban <rurban@x-ray.at>

$Id$