Hi again,
Ever curious, I decided to test the theory. I loaded the home page (of my retail site, below), then a category page and then a product page. In each case, after the page first loaded, I reloaded it twice more.
Then I used the first of the two tools mentioned by Clyde above to optimize both the stylesheet.css and the stylesheet_css_buttons.css files. (I have no others that would load.)
Each time I reloaded the page I recorded the page parse time. (You can turn that on in admin/Configuration/Logging.) Here's the results...
Code:
Home Cat. Product
1.380 4.193 1.885
0.585 1.384 0.832
0.518 1.371 0.801
(0.9152)
And after optimization...
Code:
Home Cat. Product
0.617 1.389 0.883
0.513 1.389 0.867
0.519 1.423 0.874
(0.9308)
If you discard the first line in each test because it's the initial loading into memory for many elements, including images, then you can see that there is no significant difference attributable to the css files being optimized. (I didn't purge the browser cache before running the second set of tests, so many of those elements were still in memory.)
The last number in each set, the one in brackets, is the average of the last two figures in each of the three columns above it. It indicates that the set using the optimized css file was slightly slower than the first set. I expect that is attributable to vagaries in transmission time between the server and me, but it might conceivably be due to the fact that the optimized file is not valid CSS and the browser has to work harder to parse it. But I don't know if that's the case.
These results confirm what I've read elsewhere, that such forms of optimization operate on the least problematic part of the problem, the loading into memory of the textual page elements. Server load and back-end code optimization -- the sort of thing being greatly improved in ZC v2 -- are much more likely to yield useful improvements.
YMMV
Rob
Bookmarks