I checked the pages SO with Firefox 3.5.1 and YSlow 1.4.0, and got:

(Links are to screenshots of YSlow results, not the actual pages on SO.)

You could check the entire site if you wanted to.

Please make all scores above 90 points, especially for speed-related factors.

  • I don't want to post your images to the SO imgur account without your permission, but I encourage you to do it yourself.
    – Pops
    Commented May 4, 2012 at 21:14

2 Answers 2


make sure you are testing this as an anonymous (logged out) user, since the rendering is significantly different in that case, and anon users account for 90%+ of our traffic.

Also, what Cletus said:

  • ETags are worthless
  • Content delivery networks are worthless for anyone except Google and Yahoo scale websites
  • We don't control a lot of the things YSlow is complaining about (eg, Gravatar images, references to Google Analytics .js and Quantcast .js, etc)

In short, Yahoo's Problems Are Not Our Problems

That said we did register sstatic.net and we do plan to serve up images and stuff from a cookieless domain at some point. That's about the only rational optimization left.

This is done.

  • Thank you, especially for the explanation.
    – akarnokd
    Commented Jul 19, 2009 at 7:42

YSlow, while useful, isn't a dogma. There are a few issues that will lower SO's score:

  • Gravatars images come from beyond SO's control and for practical reasons won't have far futures Expires headers and the like since the image name is just an MD5 hash of the email address and could change;
  • Yahoo marks down lack of CDN usage. Frankly, CDN usage is overrated in this regard;
  • You also lose points for not using ETags and there are circumstances where they add little value (like far futures expiry combined with versioning). In these cases you're only adding ETags to boost YSlow score and that's it; and
  • Lack of minified JS/CSS is an issue but a minor one if you use compressed output (which SO does). The way compression works is that if you compress something with no spaces versus something with lots of spaces, the second version will probably be larger but not much larger. That's information theory for you.

All in all, SO isn't committing any particularly egregious sins.

  • I agree with you mostly, but somehow the Edit page managed to get a good score. If it can be done there, it should be possible everywhere. And I knew about CDN and I added the hosts before I made the test.
    – akarnokd
    Commented Jul 18, 2009 at 14:53
  • All those scores are fairly good. Much of YSlow is about perceived performance. I've found A & B sites are both pleasing. By the way, Steve Souders has moved to Google and come up with more tools and a new book there (Even Faster Web Sites).
    – Nosredna
    Commented Jul 18, 2009 at 15:00
  • 1
    Wrt to compression, minifying it definitely helps, but using a Packer definitely doesn't help. Minified + Gzipped is the most effective/fastest. Commented Jul 18, 2009 at 15:00
  • @kd304 The edit page doesn't have any Gravatars, unlike almost every other page. Commented Jul 18, 2009 at 15:07
  • Yes, there are surprising things with SO. I made this evaluation because even if I use Chrome, I feel the pages a bit slow.
    – akarnokd
    Commented Jul 18, 2009 at 16:42
  • 1
    Actually, you lose points for USING etags, an A requires you don't use them.
    – Eric
    Commented Jul 19, 2009 at 5:27

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .