Browser performance is one of the key components to achieving a fast Web browsing experience. Much of the rendering of a Web page is dependent on the engine inside of the browser, and has become an increasingly important focus in the world of Web performance. Lindsey Simon is Front-end Developer for Google’s User experience team, and is one of the engineers pioneering the profiling and aggregating data on how Web browsers perform with the BrowserScope project, which allows developers to better understand the different aspects of how the browser performs and functions in rendering their code. Lindsey Simon will also be giving a talk on the BrowserScope project at the O’Reilly Velocity Conference 2010, in San Jose, CA.

How did you come up with the BrowserScope Project?

Lindsey Simon: Browserscope represents a combination of some backend work I was doing to store test data about browser reflow timing* and Steve Souders’ UA Profiler project. Steve had written up a series of really great Network tests along with a user agent parser. He was storing the data and rendering a real-time result table of the outcome of thousands of users taking the tests. For the most part Browserscope is just a generalization of that idea to support more test categories plus a gathering together of metrics in order to create a “Summary” category. Our motivation was definitely to try to influence how features are tackled and identified by browser vendors as well as to provide a matrix for web developers to figure out what is and isn’t working in the wild in today’s (and yesterday’s) browsers.

Note: You can still get the reflow test bookmarklet even though this is not an official Browserscope category at http://www.browserscope.org/reflow/about

What is the advantage to Web developers for profiling how various browsers perform?

Lindsey Simon: For a simple example, let’s say you run a site where all your users are on browser X – and you want to make it as fast as possible for them. One great way to make decisions about how to group and serve resources would be to look at the results in the Network category for that browser. Another example – you’ve been toiling with the details of rich text editing (which is really hard!) in the browser, you can have a look at the Browserscope compatibility table to make decisions about approaches to programming your way around inconsistencies.

What are some of the key metrics that BrowserScope identifies?

Lindsey Simon: This one’s easy – for a categorical test to make it to the homepage of Browserscope it has to come with a judgmental score from 0-100 that pertains to the value of passing or failing a given test, which means that our aggregated Summary score represents the hard work of several experts in their fields at rating a browser. The categories today are: Network, Security, Rich Text, Selectors API, and ACID3. Soon we’ll be adding a category called JSKB which measures Javascript compatibility and HTML5 and CSS3 tests as well.

How do you identify what are the important characteristics of a browser’s performance to trend?

Lindsey Simon: I go for the overall score graphs over time – these are the most fun for browsers that release often, like Firefox, Chrome, and Webkit nighty. Clicking the graph button in the results table on a page like this is fun: http://www.browserscope.org/?category=summary&v=3&ua=Firefox*

What do you enjoy most or get out of the Velocity series of conferences?

Lindsey Simon: There’s so much variety at Velocity, with performance being tackles by the front and backend experts. I am fascinated by the congruencies and differences. This is a really technical conference, which is something I appreciate too. I want the details, I want to be stumped during a talk so that I go back later and read through the slides and do my own research.

About Linsdey Simon