I only looked at the first, saw huge errors in the test, and stopped
looking. So take that for what it's worth.
The first test, test 13:
Most frameworks, such as jQuery, use a wrapper for elements. Meaning when
you use their element fetching methods, they return an object, that has the
element somewhere on an internal property. The jQuery method returns a
jQuery object, which is array-like, and has all the jQuery methods of
setting CSS, event handlers, etc. However, accessing the innerHTML property
of that object will be very cheap. So will setting it, since it's pretty
much equivalent to "var a = {}; a.innerHTML = 'asdf';".
That test isn't actually testing what it thinks it is. Its comparing
MooTools' setting of actual DOM nodes content, to setting simple properties
on wrapper elements. A better test would be to use
jQuery('#container')[0],innerHTML, or just
jQuery('#container').html('asdf').
On Mon, Apr 4, 2011 at 5:13 AM, Peter Hewat <peter.he...@gmail.com> wrote:
> Hello Mootoolers,
>
> I have come across an interesting javascript framework project focused on
> performance: http://myjs.fr/
>
> I was very surprised to discover how slow Mootools turned out to be in
> these tests compared to jQuery, ExtJS and especially this new My.js
> framework.
>
> Run these benchmarks for yourselves:
>
> http://jsperf.com/my-js-perf-tests/13
> http://jsperf.com/my-js-perf-tests/2
> http://jsperf.com/my-js-perf-tests/3
> http://jsperf.com/my-js-perf-tests/4
> http://jsperf.com/my-js-perf-tests/5
>
> Are these just silly benchmarks that don't mean anything in real life usage
> or is there something to worry about?
>
