I did only clear the chache before making each test. NOT before accessing the second page.
Since I dont know how 16 extra http request comparte to 200 KB less transfered I am not sure yet about the claim of speed optimization!
I think maybe there is some potential for SF to be looked at!!!!!!!
Do Not Combine Too Much
Combining scripts is a balance between download performance and cache performance (memory use). Download performance is the time that is required in order to download a script from the server. Cache performance is determined by how long the browser stores the script on the client, and the amount of memory that the cache requires.
Imagine a Web site that has pages named PageA.aspx and PageB.aspx. PageA.aspx references three scripts: Script1.js, Script2.js, and Script3.js. PageB.aspx references Script1.js, Script2.js, and Script4.js. The following list shows some of the possible script combinations that you can use:
Create a combined script that includes Script1.js, script2.js, and Script3.js. When the user requests PageA.aspx, the first three scripts are downloaded and cached. When the user requests PageB.aspx, Script1.js and Script2.js are downloaded again, because they are combined with Script4.js, which is needed by PageB.aspx.
Create a combined script that includes Script1.js and Script2.js. If the user requests PageA.aspx, the combined script and Script3.js are downloaded and cached. If the user then requests PageB.aspx, only Script4.js is downloaded. This causes just one extra download each for PageA.aspx and PageB.aspx.
Create a combined script that contains all four scripts. When a user requests PageA.aspx, all scripts are loaded. When a user requests PageB.aspx, PageB.aspx does not have to download any scripts, because they are already cached. However, if most users request PageA.aspx but never request PageB.aspx, the first request downloads more scripts than are required, which uses cache space unnecessarily.
Any of these solutions might be right for a given Web site, depending on the Web-site design and on the traffic patterns for the site.
As per my first post. I have a panelbar on the second page which I do not have on my first page. I assume this is why I get a different combind script resource delivered.
I just made a test and did trow an images gallery on a third page.
Page 1 -> 232 KB
Page 2 -> 244 KB (added Panelbar to page)
Page 3 -> 248 KB (added Panelabar + ImageGallery to page)
Total -> 724 KB
With CombineScript False
Page 1 -> 217 KB
Page 2 -> 11 KB
Page 3 -> 6 kb
Total -> 234 KB
So from a band with wise perspective at least every time one of you customers visits these 3 pages with CombineScript set to false you save about 0,5 MB
As for speed I must test it http requests vs. extra KB -> AND only the first page is really getting a lot more http request. the other pages get the same or 3-5 more.
EDIT; Did the test see attachment. According to fiddler2 roundtrip cost for loading the 3 pages (I did two load order 1,2,3 and 3,2,1) is faster for about 3 seconds when scripts are combinde. BUT this 3 seconds are all on the first page the next pages load about as fast as when scripts are combinde)
Load time - After the initial load of a page when not combining scripts because off less traffic Fiddler2 would report about 6-7 seconds less because of the viewer KB)
FAZIT: CombiningScripts makes your first page load faster. For any following page if other controls/widget are on the page the extra KB transfered will make your page (theoretically at least) load slower.
Looking forward to feedback on from Telerik on this claim as well
How about looking at my latest post and answer that as well?
PS: If you EnableRadScriptManager on a page it seems the scripts no longer get combinde even with CombinScripts set to true -> is that supposed to happen?