My first blog post was going to be about web typography… but then a little issue landed on my lap last week that made me reconsider. We received a complaint that one of our websites wasn’t working properly. Normally our first response would be to find out what browser was being used when the error occurred, but thankfully our complainant had included that information in his email. He was using Firefox 2.0.
The issue of which browsers to support is not a new one. The debate has been raging for a good couple of years with the glacially slow death of IE6. However, it is still a pretty serious question, especially for the public sector. Looking at our own statistics from the past six months we can see that the outright leader in terms of browser use is still Internet Explorer:
However, Internet Explorer’s figures are always misleading due to the concurrent running of various releases. In our statistics we find no less than 9 versions of Internet Explorer accessing our public website:
Admittedly, we can discount the bottom three – IE6 by any other name is still IE6. IE 999.1 is somebody’s hacked versioning, so we have no idea which version of IE this might be. Most worryingly, there were those 12 visits from IE5.5! (Firefox 2.0 was not used at all during this period!)
The most telling statistic though is the number of visits to our site from IE6. From 1 January 2012 to 5 July 2012 a tiny fraction over 3% of visits are made up of IE6 users. Microsoft themselves have basically stated that once the worldwide average drops below 1% it is effectively a dead browser and our average is three times that expiry rate. The good news is that our micro-site statistics suggest a much lower use of non-staff IE6 usage.
Internet Explorer 6 is a bit of a pain. In the early part of its life it was actually a reasonably good browser, but it became heavily entrenched in corporate systems. Web services and tools were being built purely with IE6 in mind as this was the corporate standard, almost ubiquitously. As such, it’s not quite so simple to say “we no longer support IE6”, at least not for public sector organisations.
So what do we do? As a Local Authority, we have recently made the decision to cease supporting IE6 for all public facing/consumed websites and projects. What do we mean by ceasing support though? We don’t mean our sites won’t work at all, we just mean we will no longer go out of our way to ensure our websites work exactly the same across the browser marketplace. Indeed, many of our sites still do work in IE6 and reasonably well. As long as we keep our code simple and apply our (still to be confirmed) design principles to our domain then specific IE6 support may not even be required.
The good news is that the vast majority of our visitors are using versions 8 and 9 of IE. Coupled with the fact that Microsoft have adjusted their approach for releasing Internet Explorer, adopting the same rapid release method as Chrome and Firefox, this will hopefully allow for more passive version control from a user perspective. The corporate world, if DCC is anything to go by, is also catching up and rolling out updates or at least opening up usage of Firefox or Chrome as alternatives.
This doesn’t solve our problems right now though, so we will continue to have some issues around browser support. The main issue at hand is one of proportion. At what point does it become unfeasible to continue expending resources (most notably time and money) in making sure a particular site is 100% across all browsers? Where do we draw our line in the sand?
Edit: 06 July 2012
Further to this article, I just found this infographic which compares browser usage between 2008 and 2012, alongside the evolution of screen sizes. These stats are based on information derived from the World Wide Web Consortium and are for global usage. What is very interesting is that globally over the past 12 months Internet Explorer in general has been marginalised as a whole. Only 19.5% of all web users are using IE in its various versions.
The information regarding screen sizes is also interesting. As a design principle we have been opting for a fixed width, centrally positioned layout based on the 960 Grid System. This information seems to clearly indicate that our work on responsive design need not be just about ensuring our sites appear correctly on smaller screens, but also appear correctly on much larger ones too!