Best Practice of Localizing Strings

deepaknet

I am looking at localizing the strings/messages in my web application. Applications predominantly use multiple include files with associative arrays. Is this the only approach or what would be the elegant recommended way for localizing strings in a typical PHP application?

traq

that's a good approach. very basic example:

<?php

$locale = "en_US";
$strings = get_strings( $locale );
if( !$strings ){ print "couldn't find strings!"; }
else{ print $strings['hello']; }

function get_strings( $locale ){
    $path = "{$_SERVER['DOCUMENT_ROOT']}/path/to/strings_$locale.php";
    if( file_exists( $path ) ){ return include( $path ); }
    return FALSE;
}

strings_en_US.php:

<?php
return array(
    "hello" => "Hello, World!"
);

laserlight

Have you considered the use of gettext?

Ashley_Sheridan

I've just been working on a site which does this (10 different languages, 100's of pages per language) and we store all the content in a database. You can normalise the tables out as necessary to avoid duplicated content.

traq

laserlight;11014047 wrote:
Have you considered the use of gettext?

I started reading up on that after your post.

One question that comes to mind has to do with modular applications - if you set up gettext with your application core, it seems you're forcing all of the modules to define and use their own textdomains. Anyone know if there's a simpler solution? A way to just extend the core .po/.mo files, or add a new path to the textdomain, as needed when a module is loaded...?

(Allowing the .po/.mo files to be kept with the module, and not require them to be saved in the core languages directory?)