Google Tech Talk on ‘The Importance of Front-End Performance’ and Yslow

April 20, 2009

Just watched another great google tech talk. Its subject is the set of rules the guys at Yahoo came up with to make web sites faster. These rules are presented on the book “High Performance Web Sites: Essential Knowledge for Front-End Engineers“, available here.

What have made my eyes jump off the orbits was one single number:

Wow! So, only 5% of the time to load is used generating and downloading the html document? In the study, Steve Souders also says he learned this happen in all websites – not only on big companies’ sites. The only ‘exception’ is google, on a primed cache, and even then the html download time was only about 36%.

This means something. You should not worry to optimize this 5% if you’re worried about overall user experience (at least until you’re sure you’ve got a great and fast frontend). This means we should worry more with the javascript we write and the way we deliver web pages and sttic content.

About the rules: They’re so simple you should give them a shot. It’s something you could try on a couple of hours, just to see if they’re really effective on your case. Post in the comments section if the rules haven’t worked for you.

About the tools: There’s YSlow from Yahoo, Firebug (of course), and plenty of others.

  • YSlow test your site against the set of rules mentioned above.
  • Firebug has its JavaScript profiler, which is great.
  • If you’re wondering which js lib to use on your next project, take a look at this link.

So the talk was great, and I’m gonna test the rules on a website I’ll be responsible to optimize over the next month. I’ll draw a graph of response time X implemented rule, and post it here when I’m done.


2 Responses to “Google Tech Talk on ‘The Importance of Front-End Performance’ and Yslow”

  1. About yout link to TaskSpeed and suggesting that it should weight on your decision for a JS framework, I happen to agree more with YUI team’s analysis ( ):

    “By and large, TaskSpeed may be more of a distraction than a source of information useful to the consumer. My greatest concern is that people will make a decision to choose one library over another based on which one can add a ridiculous number of event subscribers in 25 milliseconds vs. 30, ignoring more important issues like cross-browser stability, code maintainability, documentation quality, and community support.”

    • blogdogamma Says:

      Kind of agree, but just because web sites are not yet used for really performance-intensive tasks.

      Actually, 20% is a huge performance gain (taking Yahoo’s team example). More than that, on my current (slow) internet connection, jQuery 1.3.2 runs TaskSpeed test in almost 4 seconds! The others range from 2 to 3, so that’s a huge performance issue.

      Anyway, YUI team made a point when they said TaskSpeed is far away from (nowadays) reality with the number of event subscribers and stuff.

      The reason I got so emotional about this link is that I’m working with jQuery now and, although it is great functionally, it is sluggish as hell.

      So, I got the feeling that I could at least know it before. 🙂

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: