Hi,
Yes you're right, speed and memory is always an issue.
I have some tests here using different ways of building layers, that we were
trying out a while back:
http://www.dynapi.f2s.com/dynapi/tcanvas/
I haven't done any serious stress testing of the latest builds, since the
memory leak was solved, around May, I think we were happy that the memory
wasn't creeping up anymore.
I have found lots of small things that can save time and memory, but usually
using custom code., not pure Dynapi.
(One for instance is avoiding re-sizing images, put 100 stretched images on
a page, and check the ram use!)
One thing is pretty obvious to me though, Dynlayers can not be used as
building-blocks in the way originally meant, as one Dynlayer uses much more
memory than one normal <div> so the Dynlayers are best used as containers,
which can hold html content, manipulated by something more light-weight,
like the 1 k API.
(That's my personal experience)
If you have any suggestions on speed improvements, fire away.
Regarding developers, there are still quite a few known names popping up on
the list, but the two main developers, Pascal and Robert (and maybe Jordi)
are working on their own DOM-APIs, Roberts especially is a continuation of
DynAPI2 (very nice).
I think that regarding the further development of DynAPI we should
concentrate on fixing the bugs around Mac/Linux and IE6 (and others of
course) as opposed to redesigning the whole thing. DynAPI has reached a
mature stage, new technologies are taking over the candle (dom/svg) and
DynAPI2 should recognize that it's strength lies in it's legacy browser
support, as opposed to cutting edge pushing of the boundaries.
But for that strength to be of any use, it's browser support should be
pretty much bug-free - I am trying to find a few hours time to do some
serious bug-testing on mac/pc/linux , and to apply some of the many bugfixes
already registered at sourceforge.
I also have quite a few improvements to implement, like improved support of
CSS classes etc.
I seem to have switched into ramble-mode again, oh-well...
cheers,
Richard.
www.richardinfo.com
Oh, how I wish this list would reply to the list, not the poster... What's
wrong with the setup?
----- Original Message -----
From: "Raymond Irving" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Friday, September 28, 2001 5:28 PM
Subject: [Dynapi-Dev] Speed Test! Does DynAPI Qualify?
> Hello Everyone,
>
> I must salute the creators of DynAPI for creating such
> a fascinating API. I really like the object structure
> and it's easy-of-use. But there's something that I'm
> concerned about, and that's speed!
>
> What good is a pickup-minivan if it doesn't have
> sufficient horse power?
>
>
> After doing some testing on my pentium 200 (yes 200
> MHZ) using IE5.0, I've found DynAPI to be very slow
> when creating x number of layers.
>
> example:
>
> for(var a=1;a<=2;a++){
> y=y+22
> for (x=1;x<=20;x++){
> myLayer = new DynLayer(null,21*x,y,20,20,"#C0C0C0")
> DynAPI.document.addChild(myLayer)
> }
> }
>
> The above code will create 40 layers within 1-2
> seconds. The more layers you add the longer it will
> take to create another 40 layers. In other words, if
> you should have say 100 layers on screen and then try
> to add another 40, its takes a longer time to add the
> 40 layers than it would at the beginning.
> I've also noticed that a page takes a very long time
> to unload or reload if there are a lot of layers on
> the screen.
>
> The findLayers() function is just as slow when trying
> to find x number of inline layers.
>
> Layers are the heart of DynAPI. They're used in about
> 99% of the objects created for use with the API. For
> example let's say I want to create a dynamic web form
> that uses 40 Labels. With each Label using at least 2
> DynLayers it who take a normal computer a very long
> time to generate the 40 Labels (80 layers), not to
> mention Buttons and other widgets that I may use.
>
> The long and short of it is that DynAPI really needs
> to do something about the speed at which layers are
> created and the amount of browser-resources to
> maintain a layer.
>
> A true state of the art web application will use over
> 100 layers at any one time. As a result of this a few
> questions come to mind:
>
> 1) How long will it take for DynAPI to create x number
> layers?
>
> 2) How many layers can DynAPI handle before slowing
> down the creation of over layers?
>
> 3) How long will it take for DynAPI to unload a page?
>
> If any one knows of a way to speed up layer creations
> please let me know about it.
>
>
> P.S. I do hope the developers are on this list.
>
>
> Best regards,
> Raymond
>
>
> __________________________________________________
> Do You Yahoo!?
> Listen to your Yahoo! Mail messages from any phone.
> http://phone.yahoo.com
>
> _______________________________________________
> Dynapi-Dev mailing list
> [EMAIL PROTECTED]
> https://lists.sourceforge.net/lists/listinfo/dynapi-dev
>
_______________________________________________
Dynapi-Dev mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/dynapi-dev
