"qooxdoo Development" <[email protected]>

Re: qooxdoo is slow 

Hi.

> Just look how many DOM elements are created for simple button, this is 
> really not about selectors.

Of course, more DOM elements means more memory. But I would say it is not just 
normal but plainly unavoidable that complex applications take up a lot of 
memory. OTOH, more  memory shouldn't impede speed.

To give you an example: we have an app which uses probably several hundreds of 
widgets (nobody ever counted, I just made a quick estimate), and, the last time 
I looked, it creates quite a few thousands of DOM elements. However, it uses 
lazy initialization of each UI element (which is what I'd do in a rich client 
desktop app too). As such, almost no element is created when the application 
loads, and after startup elements are created only when they are needed. Once 
created, they are never discarded, unless the application is closed. While mem 
usage increases as more and more parts of the application are accessed, and 
more DOM elements are created, the speed of the application remains constant. 
Even when all parts of the application were accessed, and the browser east up 
to half a gig of mem, the speed is still the same as when the application 
starts.


> 2. Qooxdoo will be never for classic web-apps for speed/size reasons. 
> I can accept that, but the qooxdoo community have to accept that 
> qooxdoo will never spread (for me this is not problem at all). 

I don't know about that. Classic, mostly static web sites aren't that hip 
anymore. And more and more apps go web. Since qooxdoo seems to be the one and 
only beast of its kind out there, I guess it has no other chance than spread. 
Just the other day a colleague who has previously worked with JSP told me what 
she thinks of qooxdoo: you do the same thing ten times nicer and twice as fast 
in half the time, when using qooxdoo and JSON-RPC instead of JSP. From my 
experience, it's about the same with ASP.Net, struts, tapestry, JSF or 
whatever. Maybe it's half as much of a difference with wicket, but wicket has a 
much smaller widgets library. You can draw your own conclusion.


> 3. I will never post here another qooxdoo critique mail. I have 
> feeling that critique is not welcome here and I really tried to be as 
> objective as possible. 

I don't think criticism is not welcome on this list, on the contrary. But 
simply restating that qooxdoo is slow because it creates too many DOM elements, 
and not giving an example of how this affects your app's speed can't get you 
much else than people telling you how they do it and that it works fast enough 
for them.

Therefore, for me the question is still open: what exactly is it that works too 
slow in your applications? Under what circumstances? Maybe posting an example 
for the  playground would help. Although I'm not a developer of qooxdoo, I 
can't imagine the developers being able to optimize qooxdoo for speed where you 
need it unless they can reproduce it. Nor can people on this list come up with 
suggestions for your speed problem unless they understand it.

I suggested it before: you may be doing something in a different way that other 
developers using qooxdoo do it, which causes your speed problems. But unless 
you tell us about it, nobody can guess if there is indeed something you do 
differently, if it's just an application very heavy on the browser, or what 
other reason there might be for your problem.

br,

flj


------------------------------------------------------------------------------
This SF.net Dev2Dev email is sponsored by:

Show off your parallel programming skills.
Enter the Intel(R) Threading Challenge 2010.
http://p.sf.net/sfu/intel-thread-sfd
_______________________________________________
qooxdoo-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/qooxdoo-devel

Reply via email to