Text editors

Hi Thanh,

Glen used TinyMCE compressor in his test integration with code

<?php
$zlib_on = ini_get('zlib.output_compression');

if (empty($zlib_on)) {
?>
<script type="text/javascript" src="<?php echo $CFG->wwwroot ?>/lib/editor/tinymce3/tiny_mce_gzip.js"></script>
<script type="text/javascript">
tinyMCE_GZ.init({
 plugins : 'style,layer,table,save,advhr,advimage,advlink,emotions,iespell,insertdatetime,preview,media,'+
        'searchreplace,print,contextmenu,paste,directionality,fullscreen,noneditable,visualchars,nonbreaking,xhtmlxtras,moodleimage,moodlelink,dragmath,inlinepopups',
 themes : 'advanced',
 languages : 'en',
 disk_cache : true,
 debug : false
});
</script>
<?php
} else {
?>
<script type="text/javascript" src="<?php echo $CFG->wwwroot ?>/lib/editor/tinymce3/tiny_mce.js"></script>
<?php
}
?>

I have not tested if it has a big infuence to performance (compressor had some bug during my testing period) but http://wiki.moxiecode.com/index.php/TinyMCE:Compressor/PHP claims that

"TinyMCE Compressor gzips all javascript files in TinyMCE to a single streamable file. This makes the overall download size 75% smaller and the number of requests will also be reduced. The overall initialization time for TinyMCE will be reduced dramaticly if you use this script."

You can downlod php compressor from http://tinymce.moxiecode.com/download.php

We had a similar problem with Lesson back in the old old days with HTML Area - trouble is these things are mainly designed to show up once or twice per page - not 10x. At some point you have separate each instance into it's own namespace, so there is a limit to how much efficiency can be squeezed out of a stateless application. Two suggestions -

1. don't do that - keep the number of instances to a few per page.

2. if you need to edit multiple fields with a rich text editor, find another way (load them all into the same editor instance, for instance).

10 instances of Writely one one page would have a similar problem, and Google has spent millions (including making their own javascript engine) on that.

Thanks for the feedback everyone.

Internally, I have already compressed all the files (js and CSS) and have noticed that performance has improved.  This performance fix improves the perception of speed; i.e. with 10x editors, the editors appear to load more quickly, but the actual elapsed time to initialise the editor so that it can be used is actually the same.  So the compression gives the perception of improve speed, but the actual speed improvement is not signifincant.

From my testing, the time is taken by the loading of the TinyMCE toolbar and initialising the editor.  This time is a fix time for each editor instance.

I did come to the same conlcusion yesterday that one should think carefully whether a page should have 10x editors; and if not, make the necessary changes to avoid this situation.

I guess the question remains, if there is a good reason for 10x editors, is there any potential solution to this from TinyMCE (assuming the compression is done).