Should we still be supporting Internet Explorer 6?

Every couple of months the same topic of conversation comes up in the Web team office: should we still be supporting Internet Explorer 6? The answer so far has always been a resigned yes, but that may not be the case for too long.
A little history: IE6 was released on 27 August 2001, three days after Windows XP was released. Since then IE7 was released in October 2006, IE8 in March 2009 and IE9 public beta in September 2010. So, surely it’s now time to withdraw support for a browser that is over nine years old.
Bring down IE6
In 2009 .net magazine started a campaign called “Bring down IE6“.
Their mission:
The premise is simple: Internet Explorer 6 is antiquated, doesn’t support key web standards, and should be phased out. This isn’t about being anti-Microsoft, it’s about making sure that we encourage people to move to modern browsers such as IE8, Firefox, Chrome, Safari and Opera.
Case-by-case
In an article entitled “Calling time on IE6” Craig Grannell “asks designers and developers if it’s finally time to take IE6 behind the shed and shoot it”! He leaves the conclusion of the article to Web standards hero Jeffrey Zeldman:
How much longer we prop up this ageing browser must be decided on a case-by-case basis. Not every site can afford to dump it today, but the writing’s on the wall.
I think that’s a really important point because until recently the primary browser on the University’s default PC setup, that was installed on every Windows PC in the PC classrooms, was Internet Explorer 6. If we wanted our websites to be viewable and usable across the University then we had to support it, we had no option.
Supporting IE6 is a drag. As all web developers will know, you spend a couple of hours building something that works perfectly in Chrome, Firefox, Opera and Safari and then you spend twice as long again debugging it in IE6 and IE7 and IE8, which all appear to have introduced new bugs to the game. Keep calm and debug IE!
Analytics
Since the University’s default PC setup (‘standard build’) has now moved to Windows XP (and will hopefully soon move again to Windows 7) the default browser is now IE8, and s the requirement to support IE6 has now been reduced.
This is backed up by the statistics from our Google Analytics account that tracks which pages are being view most often and by which browsers.
Unsurprisingly Internet Explorer, being the default browser on our standard build PC, is the most popular browser to use to visit the University website; Apple Safari (the default browser on Apple Macs) is second. 42.5% of all visitors in the last month have used one version or another of Internet Explorer. The breakdown of which version is interesting:
- IE8: 79.8% (382,394 visits)
- IE7: 15.4% (73,944 visits)
- IE6: 4.4% (21,186 visits)
- IE9 beta: 0.29% (1,395 visits)
That means that only 1.8% of all visitors to the University website last month used IE6. But 21,186 visits is still quite a lot.
Frameworks
Adopting the Blueprint CSS framework a few years back made a considerable difference to our development time. Blueprint comes with a build-in IE hacks/workarounds stylesheet that addresses a good number of common IE5, IE6 and IE7 issues that has literally saved us hours and hours of hair-pulling.
Similarly we’re using the jQuery JavaScript framework which still supports IE6 and so makes cross-browser coding much simpler.
My view is that with such good support built-in to these frameworks for IE6 there’s really no excuse at the moment to completely drop providing a certain degree of support for IE6. The bugs are well known and the hacks are well-documented, and so finding workarounds for those that are not already contained in the framework files really doesn’t take that long to code these days.
Yahoo! graded browser support
However, it doesn’t mean that pages need to look pixel-for-pixel identical in every browser. Something that is made explicit in the Yahoo! Graded Browser Support chart:
Support does not mean that everybody gets the same thing. Expecting two users using different browser software to have an identical experience fails to embrace or acknowledge the heterogeneous essence of the Web. In fact, requiring the same experience for all users creates an artificial barrier to participation. Availability and accessibility of content should be our key priority.
Over the last two to three years I’ve used the Yahoo! GBS chart to inform the Web team about how much support we should be affording to the various browsers. IE6 is still granted A-grade support but it appears from a blog post “Graded Browser Support Update: Q4 2010” on the Yahoo! User Interface Blog that this is all about to change.
Listed among the various changes, which includes dropping A-grade support for Firefox 3.0 and initiating support for WebKit browsers on iOS and Android OS, is this:
Forecast discontinuation of A-grade coverage for Internet Explorer 6 in Q1 2011; we expect to move IE6 to the C-grade browser list as of the next update.
C-grade browsers, according to the GBS page are “identified, incapable, antiquated and rare.”
I would say that the bell is tolling for IE6 but it would appear from some corners of the Web that it has already rung out. Google has already held a Funeral for IE6 after it withdrew support for the aged browser. Microsoft sent flowers!
Conclusion
According to Google IE6 is already dead and buried, while Yahoo! are expected to degrade support for it in early 2011. Microsoft themselves, on the other hand, have committed to supporting IE6 until Windows XP SP3 support is removed in 2014; but that just means removing security issues rather than adding new features. IE6 will never, on its own, support HTML5 or CSS3, for example.
So, should we still be supporting Internet Explorer 6? I expect that we’ll follow Yahoo!’s lead next year and move to providing only a base level of support for it. When we move to using HTML5 and CSS3 then I expect we’ll have to drop support for IE6 completely.
We’ll make sure that content is readable but not worry too much about the presentation (CSS) and behaviour (JavaScript) layers; we’re already kind of doing that already in places, to be honest. But as we’re using frameworks for CSS and JavaScript which still support IE6 the elderly blue ‘e’ may be inadvertently supported for a little while to come.
Then all we need to do is try to kill off IE7. Who’s with me?
I think 1.8% is quite high unfortunately. So one eye should be kept on IE6 for the time being, even though it is an absolute pain! Worth noting is the fact that, with all the talk surrounding how important it is to start thinking about the mobile web, only 14,690 users accessed the website on a mobile device in the last month -- over 6,000 fewer users than use IE6. If supporting mobile is important, supporting IE6 is even more important. We should be honest about that fact. But I think (hope) that the figures for IE6 and mobile will have reversed by this time next year. Fingers crossed we will be able to take our eyes off IE6 soon.
Duncan, I think supporting IE6 is more of an inconvenience than an "absolute pain". Using Blueprint CSS and jQuery frameworks have downgraded it from an absolute pain, in my opinion. I don't agree with your logic that "if supporting mobile is important, supporting IE6 is even more important" based on numbers alone. I think that supporting the mobile Web (which is being actively developed) is far more important than supporting a nine year old, legacy browser that's holding on to current Web standards by the skin of its teeth. Yahoo!'s C-grade support is this:
I'd be more than happy for us to move to that level of support for the University website next year. It's pretty much what we're doing already: make sure it doesn't break terribly but don't spend too much trying to get things to line-up in a pixel-perfect way with other, more modern, more compliant browsers. As I said Blueprint CSS and jQuery certainly give us more than a helping hand in that direction. I suspect that holding on to supporting IE6 as A-grade support much longer may have more of an impact on our Web software developers than those of us concerned more with 'simple' HTML, CSS and moderate helpings of JavaScript sites.Hi Folks, There are many arguments to “Why” someone should upgrade from IE6. As with all technology, it is ever changing and improving. In the case of the Internet, we now have new modern standards, such as HTML5 and CSS3 that demand a modern browser to be able to articulate those standards. In addition to the web other factors such as Processors, in the case of 64-bit technology and operating systems require that other technologies keep pace. As important as the technology, the issue of Security should be paramount on folks mind. Modern browsers have far better security features and leave the user less vulnerable to attacks, hacking, and viruses. Then there are the “fluffy” reasons to upgrade, performance, features, and usability, that may be a vast improvement over the older versions but less important to some. In the case of IE6 it isn’t that the newer browsers are just better looking, faster or have newer features, they are somewhat required to be able to interact with the modern web environment. This doesn’t mean that they have to buy new computers or the latest operating system; in the case of XP you can install IE8 on it (for free) and have a safer, faster more rich browsing experience. The move to a newer browser such as IE9 shouldn’t be thought of a forced change but the key to opening the new Beauty of the Web. Cheers, Rick IE Outreach Team
Of course, what I didn't mention explicitly but did touch upon in my blog post is that sometimes some users cannot upgrade from IE6 on their machines, either to later versions of IE (perhaps because they are locked to still using Windows 2000 on a corporate managed-build) or to other browsers such as Firefox or Chrome because security permissions do not allow them to install software. But that does sounds to me like an endorsement from the IE Outreach Team to drop A-grade support for IE6 sooner rather than later and encourage users to upgrade. I'd be happy with that: upgrade for the sake of security and features. I am certainly looking forward to IE9 launching (even if IE9 public beta did break my localhost homepage!)