Language file, Best practice

[FFMG]

Hi,

I would like to add international support for my site to allow some of
my users to translate the site if they really feel the urge to do it.

I don't use any CMS, the site was developed, (and refined :)), by me.
But I see that various CMSs handle languages very differently.
Joomla! for example has a smallish define file that has all the
translations.
Wordress on the other hand seems to have a function __(...) that is
altogether a lot more complicated but also more flexible.

But I am not sure that the Joomela! approach is the best, defining what
could become a rather big language file seems a bit silly to me, (well
in my case it seems to be). And there does not seem to be any naming
convention of all those defines, that could spell<sictroubl e.

So, what is the best way, so far, of handling languages?

FFMG?


--

'webmaster forum' (http://www.httppoint.com) | 'webmaster Directory'
(http://www.webhostshunter.com/) | 'Recreation Vehicle insurance'
(http://www.insurance-owl.com/other/car_rec.php)
'Free URL redirection service' (http://urlkick.com/)
------------------------------------------------------------------------
FFMG's Profile: http://www.httppoint.com/member.php?userid=580
View this thread: http://www.httppoint.com/showthread.php?t=14085

Message Posted via the webmaster forum http://www.httppoint.com, (Ad revenue sharing).

[Mike P2]

Re: Language file, Best practice

On Apr 23, 2:05 am, FFMG <FFMG.2ph...@ no-mx.httppoint.co mwrote:

Hi,
>
I would like to add international support for my site to allow some of
my users to translate the site if they really feel the urge to do it.
>
So, what is the best way, so far, of handling languages?

I think phpBB and similar use files for language for portability and
ease of updating. I would save smaller strings (like messages, link
text, page titles) into a database and retrieve the needed language on
each page. I think SQL caching should be effective here because each
language should use the same query every time.

If you think it would be faster, you can save it in both the database
and into files (into a PHP file, with array assignments), and each
time it needs to be updated the form is filled with data from the
database, and then when the form is submitted the data would be
updated in the database and the language file would be regenerated
with that new data. Then the files that use the language would include
the appropriate language file instead of querying the database each
time (useful if the table gets big and the queries start taking too
long). I've benchmarked with 100 rows of data and 100 files, and found
that retrieving an indexed row from the database is about twice as
fast as reading the file named with that index. I've been told that
the database would be even faster in comparison if there were more
files and rows.

-Mike PII

[Mike P2]

Re: Language file, Best practice

I did some more tests (I am working on a multi-lingual site also so
this is useful to me too), and it looks like reading and unserializing
a serialized language array from a file is much quicker than reading
and unserializing from a database (on my server environment). The
first two times are using PHP to create a PHP file and then
include()ing it

Array
(
[create file] =0.007076025009 1553
[direct read] =0.009675025939 9414
[write serial] =0.005351066589 3555
[read serial] =0.000601053237 91504
[insert db serialized] =0.054249048233 032
[get db serialized] =0.037890911102 295
)

Here's the (PHP5) test that you can try on your server:
<?php
$db = new mysqli( 'dbhost', 'dbuser', 'dbpass' );
$db->query( 'USE `dbtable`' );

$i = 0;
$lang = array();

define( 'DATAFILE', 'benchmarking.p hp' );

//create file
$data = '';
$times = array( 'create file' =microtime( true ) );
$data = "<?php\n\$lang= array('asdfasdf asdf'";
for( $i = 20000; $i < 20500; ++$i )
$data .= ",'asdfewqrhbeq stn{$i}asdfasdf asd{$i}fasdf'";
$data .= ');?>';
file_put_conten ts( DATAFILE, $data );
$times['create file'] = microtime( true ) - $times['create file'];

unset( $data );

//direct read
$times['direct read'] = microtime( true );
require( 'benchmarking.p hp' );
$times['direct read'] = microtime( true ) - $times['direct read'];

//write serialized
$times['write serial'] = microtime( true );
file_put_conten ts( DATAFILE, serialize( $lang ) );
$times['write serial'] = microtime( true ) - $times['write serial'];

$lang = array();

//read serialized
$times['read serial'] = microtime( true );
$lang = unserialize( file_get_conten ts( DATAFILE ) );
$times['read serial'] = microtime( true ) - $times['read serial'];

unlink( DATAFILE );

$data = '';

//db serialized
$times['insert db serialized'] = microtime( true );
$data = serialize( $lang );
$sql_query = 'INSERT INTO `bm2` ( `data` ) VALUES ('
."'$data' )";
$db->query( $sql_query )
or trigger_error( $db->error, E_USER_ERROR );
$times['insert db serialized'] = microtime( true ) - $times['insert db
serialized'];

$data = '';
$lang = array();

//get db serialized
$times['get db serialized'] = microtime( true );
if( !( $sql = $db->query( "SELECT `data` FROM `bm2` WHERE `id` = $db-

>insert_id" ) ) )

trigger_error( $db->error, E_USER_ERROR );
list( $data ) = $sql->fetch_row();
$lang = unserialize( $data );
$times['get db serialized'] = microtime( true ) - $times['get db
serialized'];

print_r( $times );
?>

The table looks like:
`bm2` (collation is utf-8)
`id` PRIMARY INT NOT NULL auto_increment
`data` TEXT