Whenever the conversation turns to browsers lately, the question comes up: Can Microsoft be agile enough? Sure, what they're showing off in the IE9 platform previews now is interesting, but it's unfinished. Microsoft is planning to release an IE9 beta on September 15, but how long will it take for that beta to turn into a final product? And how quickly before the competition leapfrogs it?
Those doubts are understandable. Over the years, Microsoft hasn't exactly developed a reputation for swift, sure software development. But how does their performance compare with rival software developers? I went back and looked at the record, counting the number of days between major releases for IE, Firefox, Google Chrome, and Safari. Here's what the results look like, in chart form (click to see the full version in its own window):
I had to make a couple assumptions for this chart. I assumed that the final release of Internet Explorer would be on March 30, 2011, roughly six months after the beta and around the time of Microsoft's MIX conference. I think that's a reasonable period of time for the full beta cycle to complete. (By way of comparison, Microsoft went from beta to RTM of Windows 7 in less time than that.) Performance on the IE9 development effort has been very steady, with new releases every 6-8 weeks. So it can certainly be done. I also gave Mozilla credit for its Beta 2 release of Firefox 4 in July. Even with that largesse, they've still taken an unusually long time between major releases.
It's hard to fully gauge what Microsoft is capable of doing based on past performance. Every single version of Internet Explorer up till now has been tied to a new release of Windows, which explains the enormous gap between IE6 (Windows XP, 2001) and IE7 (Vista, 2006). Clearly, Microsoft realizes that three years might be a reasonable gap between Windows releases, but it's far too big a gap between browser updates. So what is the right number? In his keynote address at MIX06, Bill Gates was fairly blunt:
The browser we need to be unbelievably agile with. I don't know if [the proper release cycle is] nine months or 12 months or what it is, but it's much more like that than what we've done for these last three years.
Based on recent performance, Microsoft is a long ways from being able to deliver a new browser every year. Ironically, Apple is there already, releasing Safari 5 364 days after Safari 4. And Google is working at twice that speed, releasing Chrome 5 almost exactly six months after its predecessor. That's understandable, given Chrome's minimal user interface.
Picking the right release cycle is a tremendous balancing act for Microsoft, one in which they have to accommodate the demands of conservative corporate customers (who want to avoid upgrades except when absolutely required) and big-spending, trend-setting early adopters, who crave change.
The big question is whether IE9 represents a true break from the past for Microsoft. From a standards point of view, that's certainly true, and its development effort also suggests a tempo that it hasn't come close to in the past. Maybe after IE9 is complete, Microsoft will finally be able to pick up the pace, with the engine evolving along with the W3C's HTML5 specifications. If that's the case, an annual browser update could be the norm, with Internet Explorer 10 ready in early 2012, in time to be included with Windows 8.