Multlingual Websites?

quicknik

Hi,
I wonder what would be the best way of building
multilingual websites.
I would be grateful for any experience sharing.

Cheers,
nik

MarkR

If two or more of the languages do not share (or do not use) an 8-bit encoding, you should use UTF-8.

PHP has crap Unicode support, I strongly urge you not to use PHP if you need Unicode.

Assuming that your multilingual site will have several languages which share an 8-bit encoding (like Western European, English, French, Spanish, Italian, German), then it should be easy enough.

Example:
OK: English, French, Spanish, German, Italian
OK: English, Russian
Not OK: English, Russian, French (Because the French character set does not contain Russian characters and vice versa)
Really really not ok: Japanese, Chinese etc

Make a file with constants for messages, and define one copy of this file for each language. When you need to choose the language, choose one of these files, include it, then you can use it throughout.

Of course any items stored in the database need to have multilingual versions available, which may make the schema more complicated. But I'm sure you can handle that.

Mark

quicknik

Mark,

thank you so far. I guessed the same way you descriped.
But how to handle layout/HTML stuff if i do not want to do a word
by word translation.

e.g.

DE.inc.php
$HEADLINE='HALLO <b>WELT</b> ';

ENG.inc.php
$HEADLINE='HELLO <b>WORLD</b>';

index.php
include ("$LANG.inc.php");
echo $HEADLINE;

As you can see things will become a little ugly if i would decide to change the looking of the word "WORLD" for example.
Any ideas?

bye,
nik

MarkR

You need to decide how fine-grained to make it. Usually it isn't word-for-word:

<?php
  // Main file, obtain $lang somehow; cookies or user prefs, ensure they cannot inject arbitrary strings in it for security.
 // Could also use content-negotiation based on browser's accept-language header.
  require "language/$lang.php";

  echo GREETING;
?>

en.php

<?php

 define('GREETING','Hello, <b>World</b>');
?>

de.php

<?php

 define('GREETING','Hallo, <b>welt</b>');
?>

Or something

Mark

ppowell

I don't propose any of this, to be honest, simply because it becomes an issue of grammar. "Hello world" in English has a totally different grammatical structure in Finnish, Hungarian, Arabic, Farsi, Urdu, Quechua and Bushmen Click Language, hence, the idea of machine-level translation is not always a good idea.

The ideas we came up with years ago involved human-level translation using a database and a user-interfaced content management system to allow for the native speakers to grammatically compare "Hello World" - the literal phrase - to their language.

The end-user would be given a choice between viewing a content-driven website in their native language (choosing their own saved locale), the language of the environment (the server-stored locale information), or the default langauge (no locale).

The methodology to do this is totally feasible in PHP as we were able to do it entirely in TCL years earlier, and PHP lends itself moreso toward web development than TCL does.

Phil