Internet Explorer 9, web standards and Acid3

Since the MIX10 IE9 announcement last week, discussion around IE9’s standards support has been raging.

For years, Internet Explorer has been ridiculed for lousy scores on tests such as Acid3, which delivers a percentage score of standards compliance. Chrome and Opera score a 100, Firefox rolls up to 94 and IE8… Well, IE8 cranks out a measly 20. IE9 platform preview now delivers 55/100 – not bad, but really, there are mobile browsers that do better.

The real conflict doesn’t come from the Acid3 score, it comes from Microsoft’s attitude. They have publicly stated that 100/100 Acid3 score is not their goal. Neither is beating the competition in the SunSpider JavaScript benchmark. Instead, they have gathered a representative sample of what the web actually does (both in the APIs used and performance), and focus on improving the overall experience. Sounds like a Sinofskyan statistical approach familiar from the Windows 7 planning process, right?

If you want to understand Microsoft’s stance in more detail, check out the Ars Technica article on the issue. It demonstrates Microsoft’s point well – actually so well that it forgets all about critical journalism, but the comments take decent care of that part.

So how should it be?

IE9 is clearly taking a leap ahead in the web standards support. Those bashing it do make a decent point however: If Microsoft bases its set of supported standards on data it has collected, it essentially remakes the standards. Then again, as a method of setting priorities, that data collection is simply great.

If people trusted Microsoft to listen to developers and keep the browser evolving, nobody would have a problem with any of this. IE’s competitors got to set the moral playground by being more open, playing the Acid3 game and then whining about IE’s lacking standards support. Undeniably IE sucks when it comes to web standards, but does it suck because it doesn’t get a high Acid3 score or perhaps some other features entirely untested by the suite?

The MIX10 day 2 keynote makes a bold statement: user and developer experience matters more than tests. See it from minutes 13:00 to about 29:45, and you’ll see my point. The demos are actually pretty impressive. So what if IE9’s SVG, (or CSS 3, HTML 5, whatever) support isn’t exactly perfect, if its performance and delivery is so good that it actually enables a new class of applications?

Personally, I think Microsoft took the right turn here. Instead of focusing on 100% standards coverage (note, it’s different from “compliance”), focus on the things people use. Focus on getting IE9 out as a browser that actually enables the web to evolve. It is about time for the importance of piecemeal tests to be challenged by real-world applications – and yes, I agree that this might already have happened if IE hadn’t been dragging the developers down.

Microsoft can still fail – if it stops at IE9 or doesn’t listen. While people will certainly whine about IE9’s standards support, it’s still a leap ahead from where we are now. But IE10 needs to be equally revolutionary to keep things moving, and the developers need to be heard. And there needs to be a roadmap when IE9 comes out, and it must be frequently updated.

It’s a shame that the competing browser vendors and W3C couldn’t advance the web more while Microsoft was sleeping. Now that IE9 really seems to be bringing the next wave here, it'll be interesting to see the effects on web development: How will the newer specifications like HTML 5 and CSS 3 survive in the hands of the mainstream developers?

I think this is a really, really delicious moment for the web.

March 22, 2010 · Jouni Heikniemi · One Comment
Tags: IE · Posted in: Web