If it costs a large enough amount of money, they would. More precisely, they should. Why stop at IE6? There's still a combined large number of users using IE5 and IE5.5 and Lynx and NS4 and Amaya, etc.

In some cases, it may cost quite a large amount to continue to maintain and patch sites to also support IE6, and that money needs to be deducted from the 'profit' had from IE6 users. And more to the point, that same money being spent supporting those IE6 users could be being spent on bolstering mobile device work, newer browsers, assistive technologies, and what not.

The "well, we don't want to lose the profit from our users" has been thrown at me in the past, with the charge that "well, techies just care about gadgets/technology". No... in reality, many of us in the trenches (not all, but many) end up being better positioned to see the real business effects of these decisions. I'd much rather be working for a company that was spending its money wisely and being forward thinking rather than reactionary and wasteful. That's not a tech issue - it's a business issue.

Those browsers are not alike. If you follow standard advice - valid HTML, progressive enhancement, unobstrusive Javascript (or fallbacks where needed) - Lynx will render your snazzy 2011 page with the same fidelity as a 1995 Geocities page, and the result will be perfectly usable. IE6, though, will not only ignore your CSS3 stuff, but also completely misinterpret the things it does "understand".

2% may use IE6, but I'd guess that 2% uses the web WAY less than the other 98%. IE6 is probably 0.5% or less of total page views. As a group, IE6 users are also probably significantly less likely to spend money on the web.

Then there's opportunity cost. Time and budget spent on supporting IE6 could have gone into additional features. It's true that supporting IE6 is a business decision and it could be worth it for some companies, but even that 2% statistic makes IE6 seem a lot more relevant than it really is.

Opera has more market share. However, many sites don't support it. The iPad is already at 1% market share and should easily be at 2% within the next 12 months. It really depends on your target audience, of course, but which group of people would you guess has more extra cash to spend? I would target the iPad audience before the IE6 audience.

Considering so many use IE6 at work and so few at home, 2% of visits != 2% of the population in any meaningful way. So the question is more like: how important is it for my application that users at big corporations with outdated software can still use my app while at work instead of just at home?

The blog post talks about Facebook development. Has Facebook ever released browser share numbers? I bet their target audience is a lot less than 2% IE6.

I was responding to a specific comment which suggested 2% is nothing worth caring about, not to the original blog.

That said, I would think that Facebook is mainstream enough to get it's fair share of non-technical users.

IE6 usage has nothing to do with technical vs non-technical anymore. Most IE6 usage in the US is due to stupid corporate policies, home users long ago moved on, and of course the current widespread mobile browsers are all way past IE6.

I optimistically hypothesize that users are less likely to access Facebook (or most other non-internal webapps) from a locked-down corporate desktop, and even if not, that those users also access Facebook from non-corporate systems, making the loss of access from the corporate system partly illusory -- if the posited loss is "2%", I think the actual loss is probably as low as 1%.