repo_zf1/documentation/manual/en/module_specs/Zend_Translate-Plurals.xml

<?xml version="1.0" encoding="UTF-8"?>
<!-- Reviewed: no -->
<sect1 id="zend.translate.plurals">

    <title>Plural notations for Translation</title>

    <para>
        As of Zend Framework 1.9, <classname>Zend_Translate</classname> is able to provide plural
        support. Professional translation will always have the need to use plurals as they are
        native in almost all languages.
    </para>

    <para>
        So what are plurals? Generally spoken plurals are words which take into account numeric
        meanings. But as you may imaging each language has it's own definition of plurals.
        English, for example, supports one plural. We have a singular definition, for example
        "car", which means implicit one car, and we have the plural definition, "cars" which could
        mean more than one car but also zero cars. Other languages like russian or polish have
        more plurals and also the rules for plurals are different.
    </para>

    <para>
        When you want to use plurals with <classname>Zend_Translate</classname> you must not need
        to know how the plurals are defined, only the translator must know as he does the
        translation. The only information you need to have is the language.
    </para>

    <para>
        There are two way for using plurals... the traditional one, which means that you use a own
        method, and a modern one, which allows you to do plural translations with the same method
        as normal translations.
    </para>

    <sect2 id="zend.translate.plurals.traditional">

        <title>Traditional plural translations</title>

        <para>
            People who worked with gettext in past will be more common with traditional plural
            translations. There is a own <methodname>plural()</methodname> method which can be
            used for plural translations.
        </para>

        <example id="zend.translate.plurals.traditional.example1">

            <title>Example of traditional plural translations</title>

            <para>
                The <methodname>plural()</methodname> method accepts 4 parameters. The first
                parameter is the singular messageId, the second is the plural messageId and the
                third is the number or amount.
            </para>

            <para>
                The number will be used to detect the plural which has to be returned. A optional
                forth parameter can be used to give a locale which will be used to return the
                translation.
            </para>

            <programlisting language="php"><![CDATA[
$translate = new Zend_Translate('gettext', '/path/to/german.mo', 'de');
$translate->plural('Car', 'Cars', $number);
]]></programlisting>

        </example>

    </sect2>

    <sect2 id="zend.translate.plurals.modern">

        <title>Modern plural translations</title>

        <para>
            As traditional plural translations are restricted to source code using english plurals
            we added a new way for plural translations. It allows to use the same
            <methodname>translate()</methodname> for standard and for plural translations.
        </para>

        <para>
            To use plural translations with <methodname>translate()</methodname> you need to give
            an array as messageId instead of an string. This array must have the original plural
            messageId's, then the amount and at last an optional locale when your given messageId's
            are not in english notation.
        </para>

        <example id="zend.translate.plurals.modern.example1">

            <title>Example of modern plural translations</title>

            <para>
                When we want to translate the same plural definitions like in the previous our
                example would have to be defined like below.
            </para>

            <programlisting language="php"><![CDATA[
$translate = new Zend_Translate('gettext', '/path/to/german.mo', 'de');
$translate->translate(array('Car', 'Cars', $number));
]]></programlisting>

        </example>

        <para>
            Using modern plural translations it is also possible to use any language as source
            for messageId's.
        </para>

        <example id="zend.translate.plurals.modern.example2">

            <title>Example of modern plural translations using a different source language</title>

            <para>
                Let's expect we want to use russian and let's also expect that the given
                messageId's are russian and not english.
            </para>

            <programlisting language="php"><![CDATA[
$translate = new Zend_Translate('gettext', '/path/to/german.mo', 'de');
$translate->translate(array('Car',
                            'Cars first plural',
                            'Cars second plural',
                            $number,
                            'ru'));
]]></programlisting>

        </example>

        <para>
            As you can see you can give more than just the one english plural. But you must give
            the source language in this case so <classname>Zend_Translate</classname> knows which
            plural rules it has to apply.
        </para>

        <para>
            When you omit the plural language then english will be used per default and any
            additional plural definition will be ignored.
        </para>

    </sect2>

    <sect2 id="zend.translate.plurals.source">

        <title>Plural source files</title>

        <para>
            Not all source formats support plural forms. Look into this list for details:
        </para>

        <table id="zend.translate.plurals.source.supportedadapters">
            <title>Plural support</title>
            <tgroup cols="4">
                <thead>
                    <row>
                        <entry>Adapter</entry>
                        <entry>Plurals supported</entry>
                    </row>
                </thead>
                <tbody>
                    <row>
                        <entry>Array</entry>
                        <entry><emphasis>yes</emphasis></entry>
                    </row>
                    <row>
                        <entry>Csv</entry>
                        <entry><emphasis>yes</emphasis></entry>
                    </row>
                    <row>
                        <entry>Gettext</entry>
                        <entry><emphasis>yes</emphasis></entry>
                    </row>
                    <row>
                        <entry>Ini</entry>
                        <entry><emphasis>no</emphasis></entry>
                    </row>
                    <row>
                        <entry>Qt</entry>
                        <entry><emphasis>no</emphasis></entry>
                    </row>
                    <row>
                        <entry>Tbx</entry>
                        <entry><emphasis>no</emphasis></entry>
                    </row>
                    <row>
                        <entry>Tmx</entry>
                        <entry><emphasis>no</emphasis></entry>
                    </row>
                    <row>
                        <entry>Xliff</entry>
                        <entry><emphasis>no</emphasis></entry>
                    </row>
                    <row>
                        <entry>XmlTm</entry>
                        <entry><emphasis>no</emphasis></entry>
                    </row>
                </tbody>
            </tgroup>
        </table>

        <para>
            Below you can find examples of plural defined source files.
        </para>

        <sect3 id="zend.translate.plurals.source.array">

            <title>Array source with plural definitions</title>

            <para>
                An array with plural definitions has to look like the following example.
            </para>

            <programlisting language="php"><![CDATA[
array(
    'plural_0' => array(
        'plural_0 (ru)',
        'plural_1 (ru)',
        'plural_2 (ru)',
        'plural_3 (ru)'
    ),
    'plural_1' => ''
);
]]></programlisting>

            <para>
                In the above example <code>plural_0</code> and <code>plural_1</code> are the
                plural definitions from the source code. And the array at <code>plural_0</code>
                has all translated plural forms available. Take a look at the following example
                with real content and translation from english source to german.
            </para>

            <programlisting language="php"><![CDATA[
array(
    'Car' => array(
        'Auto',
        'Autos'
    ),
    'Cars' => ''
);
]]></programlisting>

            <para>
                When your translated language supports more plural forms then simply add them to
                the array below the first plural form. When your source language suppors more
                plural forms, than simply add a new empty translation.
            </para>

        </sect3>

        <sect3 id="zend.translate.plurals.source.csv">

            <title>Csv source with plural definitions</title>

            <para>
                A csv file with plural definitions has to look like the following example.
            </para>

            <programlisting language="php"><![CDATA[
"plural_0";"plural_0 (ru)";"plural_1 (ru)";"plural_2 (ru)";"plural_3 (ru)"
"plural_1";
]]></programlisting>

            <para>
                All translated plural forms have to be added after the first plural of the source
                language. And all further plural forms of the source language have to be added
                below but without translation. Note that you must add a delimiter to empty
                source plurals.
            </para>

        </sect3>

        <sect3 id="zend.translate.plurals.source.gettext">

            <title>Gettext source with plural definitions</title>

            <para>
                Gettext sources support plural forms out of the box. There is no need for adoption
                as the <filename>*.mo</filename> file will contain all necessary data.
            </para>

            <note>

                <para>
                    Note that gettext does not support the usage of source languages which are not
                    using english plural forms. When you plan to use a source language which
                    supports other plural forms like russian for example, then you can not use
                    gettext sources.
                </para>

            </note>

        </sect3>

    </sect2>

</sect1>
<!--
vim:se ts=4 sw=4 et:
-->