<?xml version="1.0" encoding="UTF-8"?>
<!-- Reviewed: no -->
<sect1 id="zend.translate.adapter">
    <title>Adapters for Zend_Translate</title>

    <para>
        <classname>Zend_Translate</classname> can handle different adapters for translation.
        Each adapter has its own advantages and disadvantages.
        Below is a comprehensive list of all supported adapters for
        translation source files.
    </para>

    <table id="zend.translate.adapter.table">
        <title>Adapters for Zend_Translate</title>

        <tgroup cols="3">
            <thead>
                <row>
                    <entry>Adapter</entry>
                    <entry>Description</entry>
                    <entry>Usage</entry>
                </row>
            </thead>

            <tbody>
                <row>
                    <entry>Array</entry>
                    <entry>Use <acronym>PHP</acronym> arrays</entry>
                    <entry>Small pages; simplest usage; only for programmers</entry>
                </row>

                <row>
                    <entry>Csv</entry>
                    <entry>Use comma separated (*.csv/*.txt) files</entry>

                    <entry>
                        Simple text file format; fast; possible problems with unicode characters
                    </entry>
                </row>

                <row>
                    <entry>Gettext</entry>
                    <entry>Use binary gettext (*.mo) files</entry>
                    <entry>GNU standard for linux; thread-safe; needs tools for translation</entry>
                </row>

                <row>
                    <entry>Ini</entry>
                    <entry>Use simple <acronym>INI</acronym> (*.ini) files</entry>

                    <entry>
                        Simple text file format; fast; possible problems with unicode characters
                    </entry>
                </row>

                <row>
                    <entry>Tbx</entry>
                    <entry>Use termbase exchange (*.tbx/*.xml) files</entry>

                    <entry>
                        Industry standard for inter application terminology strings;
                        <acronym>XML</acronym> format
                    </entry>
                </row>

                <row>
                    <entry>Tmx</entry>
                    <entry>Use tmx (*.tmx/*.xml) files</entry>

                    <entry>
                        Industry standard for inter application translation; <acronym>XML</acronym>
                        format; human readable
                    </entry>
                </row>

                <row>
                    <entry>Qt</entry>
                    <entry>Use qt linguist (*.ts) files</entry>

                    <entry>
                        Cross platform application framework; <acronym>XML</acronym> format; human
                        readable
                    </entry>
                </row>

                <row>
                    <entry>Xliff</entry>
                    <entry>Use xliff (*.xliff/*.xml) files</entry>

                    <entry>
                        A simpler format as <constant>TMX</constant> but related to it;
                        <acronym>XML</acronym> format; human readable
                    </entry>
                </row>

                <row>
                    <entry>XmlTm</entry>
                    <entry>Use xmltm (*.xml) files</entry>

                    <entry>
                        Industry standard for <acronym>XML</acronym> document translation memory;
                        <acronym>XML</acronym> format; human readable
                    </entry>
                </row>

                <row>
                    <entry>Others</entry>
                    <entry>*.sql</entry>
                    <entry>Different other adapters may be implemented in the future</entry>
                </row>
            </tbody>
        </tgroup>
    </table>

    <sect2 id="zend.translate.adapter.decision">
        <title>How to decide which translation adapter to use</title>

        <para>
            You should decide which Adapter you want to use for
            <classname>Zend_Translate</classname>. Frequently, external criteria such as a project
            requirement or a customer requirement determines this for you, but if you are in
            the position to do this yourself, the following hints may simplify
            your decision.
        </para>

        <note>
            <para>
                When deciding your adapter you should also be aware of the used
                encoding. Even if Zend Framework declares UTF-8 as default
                encoding you will sometimes be in the need of other encoding.
                <classname>Zend_Translate</classname> will not change any encoding which is defined
                within the source file which means that if your Gettext source
                is build upon ISO-8859-1 it will also return strings in this encoding
                without converting them. There is only one restriction:
            </para>

            <para>
                When you use a <acronym>XML</acronym> based source format like TMX or XLIFF you must
                define the encoding within the <acronym>XML</acronym> files header because
                <acronym>XML</acronym> files without defined encoding will be treated as UTF-8 by
                any <acronym>XML</acronym> parser by default. You should also be aware that actually
                the encoding of <acronym>XML</acronym> files is limited to the encodings supported
                by <acronym>PHP</acronym> which are UTF-8, ISO-8859-1 and US-ASCII.
            </para>
        </note>

        <sect3 id="zend.translate.adapter.array">
            <title>Zend_Translate_Adapter_Array</title>

            <para>
                The Array Adapter is the Adapter which is simplest to use for
                programmers.
                But when you have numerous translation strings or many
                languages you should think about another Adapter.
                For example, if you have 5000 translation strings,
                the Array Adapter is possibly not the best choice for you.
            </para>

            <para>
                You should only use this Adapter for small sites with a handful
                of languages, and if you or your programmer team creates the
                translations yourselves.
            </para>
        </sect3>

        <sect3 id="zend.translate.adapter.csv">
            <title>Zend_Translate_Adapter_Csv</title>

            <para>
                The Csv Adapter is the Adapter which is simplest to use for
                customers.
                CSV files are readable by standard text editors, but
                text editors often do not support utf8 character sets.
            </para>

            <para>
                You should only use this Adapter if your customer wants to do
                translations himself.
            </para>

            <note>
                <para>
                    Beware that the Csv Adapter has problems when your Csv files are encoded
                    differently than the locale setting of your environment. This is due to a Bug of
                    <acronym>PHP</acronym> itself which will not be fixed before
                    <acronym>PHP</acronym> 6.0 (http://bugs.php.net/bug.php?id=38471). So you should
                    be aware that the Csv Adapter due to <acronym>PHP</acronym> restrictions is not
                    locale aware.
                </para>
            </note>
        </sect3>

        <sect3 id="zend.translate.adapter.gettext">
            <title>Zend_Translate_Adapter_Gettext</title>

            <para>
                The Gettext Adapter is the Adapter which is used most
                frequently. Gettext is a translation source format which was
                introduced by GNU, and is now used worldwide.
                It is not human readable, but there are several freeware tools
                (for instance, <ulink url="http://sourceforge.net/projects/poedit/">POEdit</ulink>),
                which are very helpful. The <classname>Zend_Translate</classname> Gettext Adapter is
                not implemented using <acronym>PHP</acronym>'s gettext extension.
                You can use the Gettext Adapter even if you do not have
                the <acronym>PHP</acronym> gettext extension installed.
                Also the Adapter is thread-safe and the <acronym>PHP</acronym> gettext extension
                is currently not thread-safe.
            </para>

            <para>
                Most people will use this adapter.
                With the available tools, professional translation is
                very simple. But gettext data are is stored in a
                machine-readable format, which is not readable without tools.
            </para>
        </sect3>

        <sect3 id="zend.translate.adapter.ini">
            <title>Zend_Translate_Adapter_Ini</title>

            <para>
                The Ini Adapter is a very simple Adapter which can even be used
                directly by customers.
                <acronym>INI</acronym> files are readable by standard text editors, but
                text editors often do not support utf8 character sets.
            </para>

            <para>
                You should only use this Adapter when your customer wants to do translations
                himself. Do not use this adapter as generic translation source.
            </para>

            <warning>
                <title>Regression in PHP 5.3</title>

                <para>
                    Prior to <acronym>PHP</acronym> 5.3, <methodname>parse_ini_file()</methodname>
                    and <methodname>parse_ini_string()</methodname> handled non-ASCII characters
                    within <acronym>INI</acronym> option keys worked without an issue. However,
                    starting with <acronym>PHP</acronym> 5.3, any such keys will now be silently
                    dropped in the returned array from either function. If you had keys utilizing
                    UTF-8 or Latin-1 characters, you may find your translations no longer work when
                    using the <acronym>INI</acronym> adapter. If this is the case, we recommend
                    utilizing a different adapter.
                </para>
            </warning>
        </sect3>

        <sect3 id="zend.translate.adapter.tbx">
            <title>Zend_Translate_Adapter_Tbx</title>

            <para>
                The Tbx Adapter is an Adapter which will be used by customers
                which already use the TBX format for their internal translation
                system. Tbx is no standard translation format but more a collection
                of already translated and pre translated source strings. When you
                use this adapter you have to be sure that all your needed source
                string are translated.
                TBX is a <acronym>XML</acronym> file based format and a completely new format.
                <acronym>XML</acronym> files are human-readable, but the parsing is not as fast
                as with gettext files.
            </para>

            <para>
                This adapter is perfect for companies when pre translated
                source files already exist.
                The files are human readable and system-independent.
            </para>
        </sect3>

        <sect3 id="zend.translate.adapter.tmx">
            <title>Zend_Translate_Adapter_Tmx</title>

            <para>
                The Tmx Adapter is the Adapter which will be used by most
                customers which have multiple systems which use the same
                translation source, or when the translation source must be
                system-independent.
                TMX is a <acronym>XML</acronym> file based format, which is announced to be the
                next industry standard.
                <acronym>XML</acronym> files are human-readable, but the parsing is not as fast
                as with gettext files.
            </para>

            <para>
                Most medium to large companies use this adapter.
                The files are human readable and system-independent.
            </para>
        </sect3>

        <sect3 id="zend.translate.adapter.qt">
            <title>Zend_Translate_Adapter_Qt</title>

            <para>
                The Qt Adapter is for all customers which have TS files as their
                translation source which are made by QtLinguist.
                QT is a <acronym>XML</acronym> file based format.
                <acronym>XML</acronym> files are human-readable, but the parsing is not as fast
                as with gettext files.
            </para>

            <para>
                Several big players have build software upon the QT framework.
                The files are human readable and system-independent.
            </para>
        </sect3>

        <sect3 id="zend.translate.adapter.xliff">
            <title>Zend_Translate_Adapter_Xliff</title>

            <para>
                The Xliff Adapter is the Adapter which will be used by most customers which
                want to have <acronym>XML</acronym> files but do not have tools for TMX.
                XLIFF is a <acronym>XML</acronym> file based format, which is related to TMX but
                simpler as it does not support all possibilities of it.
                <acronym>XML</acronym> files are human-readable, but the parsing is not as fast
                as with gettext files.
            </para>

            <para>
                Most medium companies use this adapter.
                The files are human readable and system-independent.
            </para>
        </sect3>

        <sect3 id="zend.translate.adapter.xmltm">
            <title>Zend_Translate_Adapter_XmlTm</title>

            <para>
                The XmlTm Adapter is the Adapter which will be used by customers
                which do their layout themself. XmlTm is a format which allows the
                complete <acronym>HTML</acronym> source to be included in the translation source, so
                the translation is coupled with the layout.
                XLIFF is a <acronym>XML</acronym> file based format, which is related to XLIFF but
                its not as simple to read.
            </para>

            <para>
                This adapter should only be used when source files already exist.
                The files are human readable and system-independent.
            </para>
        </sect3>
    </sect2>

    <sect2 id="zend.translate.adapter.selfwritten">
        <title>Integrate self written Adapters</title>

        <para>
            <classname>Zend_Translate</classname> allows you to integrate and use self written
            Adapter classes. They can be used like the standard Adapter classes which
            are already included within <classname>Zend_Translate</classname>.
        </para>

        <para>
            Any adapter class you want to use with <classname>Zend_Translate</classname> must be a
            subclass of <classname>Zend_Translate_Adapter</classname>.
            <classname>Zend_Translate_Adapter</classname> is an abstract class which already defines
            all what is needed for translation. What has to be done by you, is the definition of the
            reader for translation datas.
        </para>

        <para>
            The usage of the prefix "Zend" should be limited to Zend Framework.
            If you extend <classname>Zend_Translate</classname> with your own adapter, you should
            name it like "Company_Translate_Adapter_MyFormat". The following code shows an
            example of how a self written adapter class could be implemented:
        </para>

        <programlisting language="php"><![CDATA[
try {
    $translate = new Zend_Translate(
        array(
            'adapter' => 'Company_Translate_Adapter_MyFormat',
            'content' => '/path/to/translate.xx',
            'locale'  => 'en',
            'myoption' => 'myvalue'
        )
    );
} catch (Exception $e) {
    // File not found, no adapter class...
    // General failure
}
]]></programlisting>
    </sect2>

    <sect2 id="zend.translate.adapter.caching">
        <title>Speedup all Adapters</title>

        <para>
            <classname>Zend_Translate</classname> allows you use internally
            <classname>Zend_Cache</classname> to fasten the loading of translation sources. This
            comes very handy if you use many translation sources or extensive source formats like
            <acronym>XML</acronym> based files.
        </para>

        <para>
            To use caching you will just have to give a cache object to the
            <methodname>Zend_Translate::setCache()</methodname> method. It takes a instance of
            <classname>Zend_Cache</classname> as only parameter. Also if you use any adapter direct
            you can use the <methodname>setCache()</methodname> method. For convenience there are
            also the static methods <methodname>getCache()</methodname>,
            <methodname>hasCache()</methodname>, <methodname>clearCache()</methodname> and
            <methodname>removeCache()</methodname>.
        </para>

        <programlisting language="php"><![CDATA[
$cache = Zend_Cache::factory('Core',
                             'File',
                             $frontendOptions,
                             $backendOptions);
Zend_Translate::setCache($cache);
$translate = new Zend_Translate(
    array(
        'adapter' => 'gettext',
        'content' => '/path/to/translate.mo',
        'locale'  => 'en'
    )
);

// to clear the cache somewhere later in your code
Zend_Translate::clearCache();
]]></programlisting>

        <note>
            <para>
                You must set the cache <emphasis>before</emphasis> you use or initiate
                any adapter or instance of <classname>Zend_Translate</classname>. Otherwise your
                translation source will not be cached until you add a new source with the
                <methodname>addTranslation()</methodname> method.
            </para>
        </note>

        <para>
            When the attached cache supports tagging you can set a own tag string by using the
            option <property>tag</property>. This allows you do delete only the cache from this
            single instance of <classname>Zend_Translate</classname>. When you are not using this
            option the default tag <classname>Zend_Translate</classname> is used.
        </para>

        <para>
            Using the option <property>tag</property> you must give the used tag to
            <methodname>clearCache()</methodname> to declare which tag you want to delete.
        </para>

        <programlisting language="php"><![CDATA[
$cache = Zend_Cache::factory('Core',
                             'File',
                             $frontendOptions,
                             $backendOptions);
Zend_Translate::setCache($cache);
$translate = new Zend_Translate(
    array(
        'adapter' => 'gettext',
        'content' => '/path/to/translate.mo',
        'locale'  => 'en',
        'tag'     => 'MyTag'
    )
);

// somewhere later in your code
Zend_Translate::clearCache('MyTag');
]]></programlisting>
    </sect2>
</sect1>
<!--
vim:se ts=4 sw=4 et:
-->