Google Analytics has announced it will end IE8 support by the end of the year, following Google Apps, which ended support for the browser last November.

Legacy browser support remains one of the hardest problems in web development. For years, Internet Explorer 6 was a bugbear, because enterprise applications were written with it in mind. Sadly, the same is true of its descendent: nobody uses IE8 on the weekend, which means that it's probably forcibly installed on enterprise networks, where users aren't allowed to install their own software.

Internet Explorer lock-in is rife in the enterprise, because of the browser's non-standard web support and ubiquity on Windows computers. Faced with supporting IE8 or web standards as they were actually specified, many enterprise vendors went with IE8, because that's where the customers were.

Compounding the problem, IE8 is the last browser in its line that will run on Windows XP, which is still prevalent in enterprise environments (even if users are slowly making the migration to Windows 7). In other words, to run a better version of Internet Explorer, enterprise IT departments don't just have to give permission for it to be installed; they must upgrade their computers from another operating system first. This is a significant expense.

In the web development community, it's easy to be dismissive and say that these organizations should be running Linux, and shouldn't have got themselves into this situation to begin with. (I've heard this attitude a lot.) That ignores the much broader context that Windows enterprise computing sits in, including the software ecosystem and the support infrastructure that's grown up around it. Most importantly, though, if we want to sell to a customer, it's probably a good idea to support the platforms that they actually use. The larger and more security-conscious the customer, the more reticent they may be to upgrade their platform software more regularly.

So how do you balance the fact that so many customers are on Windows XP with the fact that Internet Explorer 8 is a hideous, insecure platform that must be developed for separately?

One option is to gently suggest Firefox or Chrome, which both work with Windows XP SP2. At latakoo, we'll be doing that increasingly less gently; we've already communicated to our customers that we'll be slowly phasing out support, and we'll soon be adding some visible messaging urging them to switch browsers. However, the pragmatic reality is that many users can't switch, because of their IT rules, and often because of the IE8-specific in-house apps they're running, so we can't simply turn off support, even though maintaining IE8-only code costs us extra.

Moving away from IE8 will be more secure for every organization. (Microsoft is ending support for Windows XP in 2014.) Until then, if you're an enterprise IT manager, I recommend encouraging a two-browser solution: IE8 for the apps that really need it, and a secure, modern browser for everything else (including latakoo).

For developers, there's a lot to be said for increasingly less-subtle messaging explaining why Internet Explorer 8 is a bad choice. You're providing useful advice, while also encouraging your customers to get better value for money out of your service (because more developer time can go into new and more resilient features rather then legacy browser support). But don't switch off support completely - not quite yet at least - lest you leave some of your most important customers out in the cold.