Internet Explorer 9 is released: should you switch?

onight, Microsoft officially released Internet Explorer 9 to the public. If you’re eager for details on how to download and install the final bits, I’ll have them shortly. In this post, I’ll talk about what’s in the final release

Although only five weeks have passed between the Release Candidate and the final version available for download now, that time was well spent. Several bugs I had logged in the RC are fixed, and performance is noticeably snappier. And, somewhat surprisingly, two features that weren’t available in the RC made it into the final release:

• IE 9 now supports the Do Not Track header that Mozilla proposed earlier this year. The feature is in addition to the much more activeTracking Protection feature that blocks third-party tracking sites. According to Rob Mauceri, Principal Group Program Manager for Internet Explorer, Do Not Track is implemented as a header and as part of the Document Object Model (DOM) API. As a result, says Mauceri, “sites can detect a user’s intention not to be tracked.”
• The Tracking Protection feature now covers ActiveX controls—the most popular of which is Adobe Flash. In the final version of Internet Explorer 9,  requests from ActiveX controls go through the same Tracking Protection Lists that govern websites. If you’ve blocked a third-party site, it will be blocked for access in Flash and other ActiveX controls as well. I’ll need to test this feature to see if it has an impact on local shared objects–aka “Flash cookies.”

If you’ve already downloaded and installed the IE 9 Release Candidate you probably won’t notice any changes. The core of the IE 9 interface is the same as I described last month in my in-depth look at the Release Candidate.

In a follow-up post, I’ll look at how IE 9 compares with its archrivals, Google Chrome and Mozilla Firefox. Here’s the executive summary, broken down into four key categories.

Performance: “screamingly faster”

Last week, my colleague Steven J. Vaughan-Nichols benchmarked the performance of Google’s just-released Chrome 10 and called it“screamingly fast.” That conclusion was based on a set of benchmarks that mistakenly compared the unoptimized 64-bit version of Internet Explorer 9 to the 32-bit version of Chrome. When he re-ran the tests, IE 9 came out ahead. I guess that makes the new Internet Explorer “screamingly faster” and, at least for now, dethrones Chrome as the speed king.

My experience with the final release of IE 9 supports that conclusion. In tests on two high-end desktop systems using the SunSpider benchmark, IE 9 came out on top of the just-released Google Chrome 10 by a minimum of 11%. The release candidate of Firefox 4 actually outperformed Chrome ontwo of three systems but was still 12.6% slower than IE 9. On a notebook using an ultra-low-voltage Core 2 Duo processor, the difference was even more profound: Firefox 4 and Chrome 10 needed 24% and 29% more time to finish the benchmark than IE 9.

The most important measure of performance, of course, is the real world. So I set up a test bed with browsers arranged side by side on PCs with similar configurations and found the differences between the three leading browsers to be minimal in everyday use. On graphically intensive pages, IE 9 was often able to finish loading a page faster than Chrome. The difference was especially noticeable on systems using older GPUs. The only significant differences I found were on sites that use cutting-edge HTML5 features, where the difference came down to differences in how each browser handles an emerging standard.