Strategy for supporting unicode & multi language in PHP5

Question

I have heard that PHP6 will natively support unicode, which will hopefully make multi-language support much easier. However, PHP5 has pretty weak support for unicode and mul

Answer 1

Well for app development in PHP I use CodeIgniter which takes care of handling multiple language files. It's very powerful and easy to use.

Here is a link to their Language Class

Answer 2

For translations, you can either use a framework, or just roll your own library. You can store translations in csv files and use PHP's fgetcsv() to parse it. CSV files can be edited with any spreadsheet app.

For an example, you can look at the code of Zend_Translate (part of Zend Framework). It's easy to follow along.

Answer 3

Related to usage of mb_* set of functions, at the same time of maintaining compatibility, see the mb_string.overload php.ini directive.

It will allow you to use the regular string functions which have been overloaded by the multi-byte enabled ones.

Answer 4

It's not all that hard really, but you may want to make your question a bit more specific.

If you're talking to a database, make sure your database stores data in UTF-8 and the connection to your database is in UTF-8 (a common pitfall). Make sure to run this when establishing a connection:

mysql_set_charset('utf8');

For user input, set the accept-charset attribute on your forms.

<form accept-charset="utf-8">

Serve your sites with an appropriate HTTP header:

header('Content-Type: text/html; charset=utf-8');

or at least set appropriate meta tags for your site:

<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />

Keep your source code files encoded in UTF-8.

If you keep everything in UTF-8, you usually don't need to worry about anything. It's only getting problematic once you start mixing encodings throughout your app.

If you're starting to talk about string manipulation of course, you'll have to take a little more care. Mostly you'll want to use the mb_ set of string functions, as you pointed out yourself.