[OpenLayers-Users] WFS Layers blocks excution?

[Jackey Cheung]

I've been using Firebug for some time, and do tracked OL with FB. The most 
time consumer is XML process on the FireFox. And I've done the same thing in 
IE with Visual Studio, and found out XML process consumes most process time 
too. Currently, I've made some change to the WFS layer, to utilize non-XML 
methods, for more performance. Well, deviated from "WFS" actually.

Just curious, how others do it? I haven't got time to study other similar 
maps on the map, but they seem to be better in performance. I've though a 
little about using timed process to deal with timely process, such as WFS, 
so the browser won't hang up while busy on processing XML data.


----- Original Message -----
From: "Christopher Schmidt" <crschmidt at metacarta.com>
To: "Jackey Cheung" <cheung.jackey at gmail.com>
Cc: <users at openlayers.org>
Sent: Wednesday, June 06, 2007 09:51
Subject: Re: [OpenLayers-Users] WFS Layers blocks excution?

> On Wed, Jun 06, 2007 at 09:13:49AM +0800, Jackey Cheung wrote:
>> I've updated to 2.4, it do improve quite a lot, but I'm still suffering
>> noticeable (little, but still noticeable) lag while panning, and it takes
>> around 1 second to load the layer when refresh. Is this normal?
>
> I would say you're probably dealing with pretty standard behavior at
> this point, yes.  If you get Firebug (debugging extension for Firefox)
> you can watch the time that requesting data from the server actually
> takes: this may help you determine whether your 1 second refresh time is
> server or clientside.
>
>> You've mentioned about pathological dataset. I'm using MapServer as the 
>> WFS
>> server, returns 3 data fields (ID, Name, coord) in GML. Isn't it general?
>> If you were referring to the server side process, I just wondering, does
>> server side process affect the execution flow of the client side?
>
> The server side execution doesn't affect the client.
>
>> What should I focus on to tune this WFS thing in OpenLayers?
>
> First, get firebug and time the requests.
> Second, turn on profiling and check out the time to do things in
> Firefox. (You can't do this in IE, but the performance is going to be
> mostly the same -- issues generally show up in both.) If you're seeing
> major differences between IE and Firefox, my best advice is basically
> "Good luck". I don't have any experience or advice in debuggin speed
> issues in IE.
>
> Regards,
> -- 
> Christopher Schmidt
> MetaCarta