When IE6 was launched some time in 2001, I was just a college freshman with a big interest in the Internet, a small interest in how websites work and no interest whatsoever in browsers, standards and all the other stuff that defined how things work the way they did (or at least should). As almost every boy over the course of his childhood tries to disassemble a toy to see what makes the wheels on a toy car spin, I was certainly no exception. But in my case this exploring spirt hasn't faded out yet. That's how HTML and I first met :)

After I constructed several „Hello world“ dummies, I think it was in 2003 that I launched my first real web site. The site looked fairly cool, at least from my perspective and for my level of expertise at that time. There was a whole lot of DHTML which was a big buzz back then... However, the most important thing was – it worked. At least I thought so, since I had no reason to believe otherwise. Back then I had no idea that anything other than IE5.5+ existed (anything important, that is). And even if it existed, I thought in my blessed ignorance, how different can a web page look in another browser? You've seen one, you've seen them all... The only difference is in the company that makes it and maybe GUI skin. Yeah, right... Until one day I saw my site in Netscape Navigator. The horror I saw that day is still haunting me from time to time... Javascript errors, hidden menus jumping all over the place... Horrible.

My experience back then is probably how many client-side developers feel today after testing their web apps in IE6, the very same ones they've just completed to be W3C compliant and work flawlessly in standards-compliant browsers. However, NN was at that time a tired old browser at the sunset of it's existence, and today IE6 is a big boring yet vigorous old hag that just keeps annoying you day and day again.
Due to its lack of support for advanced CSS and XHTML, IE6 can be considered to be really holding back the web. With so many alternatives available, and after the release of Internet Explorer 8 which put IE in the standards-compliancy era, it's time to take the web to the next step.

Four months after our launch, although IE6 usage on Bumblehood is still not low enough to be considered irrelevant, what's encouraging is the fact that both IE general share and IE6 share are dropping constantly at a pace greater than 0.5% on amonthly basis. This pace is a good signal, given the fact that IE6 already dropped significantly in 2007/2008 and that the last few pounds are the hardest to lose. What's discouraging enough is the fact that at this pace, it would take almost two more years for IE6 to drop to levels where it could no longer be considered worth supporting. That means two more years of excessive development needed to enable your web-apps for the uninformed minority. Two more years... Probably even Microsoft itself will abandon any kind of support for IE6.

A decision to stop supporting almost a fifth of the audience (IE6 market share accounts up to 16.94% in May 2009 according to NetApplications) is something designers and developers are not quite comfortable with. Yet this decision has to be made, whether several weeks of lower traffic (and consequentially a lower income) are a small sacrifice for a greater goal of getting the Internet out of the stone age. Having in mind that Pareto's principle can also be recognized in tweaks and hacks required for sites to work in IE6, it's not that hard to calculate what limited support for IE6 can contribute to extra development time.
When decomposed to a logical level, it is really easy – people are using IE6 and aren't considering replacing it with more up-to-date browsers simply due to the fact that „everything works“ in IE6. If „everything works“, it can't be that bad after all, so why bother changing it. As long as the websites are tweaked and hacked to work in IE6, people will use it to visit websites. Once the people are unable to use the majority of web services with IE6, its usage will be abandoned. What one in client-side development can do is inform the visitor that his experience of the website could improve significantly with a modern browser (or even going maybe a bit down the extreme road and letting people know that the web cannot be viewed in IE6), people will consider alternatives. Just imagine if Google, Gmail and FaceBook decided that IE6 is no longer supported by their apps - would that be a large enough impulse for the significant part of IE6 users to switch to alternative browsers?
What about us? As much as I'd like to completely abandon support for IE6, I'm quite sure the management will disagree with this idea, as it means 15% lower traffic. 15% for a start-up is worth pure gold, I know. Since Bumblehood already supports IE6 in read-mode beta, there's no point in changing this approach overnight. But with launch-day approaching for edit-mode beta and IE6 being unsupported in the first iteration, I sincerely hope that our stand towards IE6 will not change significantly more than providing limited support for content editing in IE6. That's the least we can do for the web.

Posted by Dino Ursic on June 30th, 2009 --- Permalink --- Tagged in categories: Software team --- Comments on Forum