My own site: http://www.ur-upload.de got 88 of 100 points^^

The google speed test highly considers the server network speed than the site optimization. To put it simple, demo and dev sites are on complete different hosting companies and datacenters and the results of the test are:

demo.cheverereto.com scores 28 (out of 100)
dev.chevereto.com scores 53 (out of 100)

And it's the exact same thing in terms of the script. So, yes... This test is quite irrelevant.

BTW, Chevereto's default theme doesn't use sprites because is intended to be a easy to understand default theme, so you can make your own themes, make your own sprites etc.

I must say that this is not correct.

Google Page Speed is not measuring the server speed/network speed.

It is only about Front End Performance Optimizing. If you want to know more, read e.g. here:

http://www.cookielessdomain.com/blog/20 … ptimizing/

http://ezinearticles.com/?Front-End-Per … id=5158202

or on the Google Page Speed website:

http://code.google.com/intl/en-EN/speed/page-speed/

Google Page Speed suggests different techniques like minifying/combining JS/CSS, using CSS sprites, ...

It is like YSlow. It can not test server performances.

Sorry to say, but not true.

I am into front end performance optimizing all day.

I know the effects when you e.g. do not combine all your javascripts into one file.
If you are not doing it... you have more http requests.

Same with graphics, css and every stuff your browser needs to load.

Basically you need to do the following:

minify/combine css
put css into head
minify/combine javascripts
put javascript into bottom
use css sprites
optimize graphic files with e.g. smushit

That's the basics. Now you could also use a cookieless domain to put your static files on.
Saves you some bytes, too.

You can also use a webserver with mod_deflate, mod_headers and mod_expires to speed up a little more.

For me this discussion ends now, and I will tell you why.

I did several tests with different websites/portals I have coded the last 2 years.

Even a university professor from germany called me in that front end performance stuff and asked for my advice.

If you think you are right you should read all my posts on cookielessdomain.com and think about how browsers load/execute ...

But thank you anyway.

You did a good job on chevereto, that's why I bought it :-)

Ofcoures valid HTML is very important.

And those "techniques" are not "copypasted" techniques...

Beside that there is much more you should have in mind when doing front end performance optimizing... but hey, you are the god here... and you are right, i am not.

have a nice evening

I never mention valid HTML because that is not a direct measure of improvement. Why? make standard HTML is like writing with good orthography and all the necessary things, is the necessary base. But, if you aren't concise, if you overload the rendering (HTML/CSS/JS) standars doesn't means nothing.

Dude... I'm not saying that I'm right. But you haven't really read my post. Even with the google thing you just keep saying that I'm wrong... Even with clear evidence.

Anyway, if you are ok doing all that fancy optimization things go ahead. I'm just saying that is a waste of time. Like make the HTML code 100% valid.