How militant some developers have become about ditching ie6 doesn't really make sense. It decision seems pretty clear: if you are a site for techy users you don't need to bend over backwards to support it, and you can probably drop it; if you are a general-use site or site that plans on having 30+ year-olds you better have some support for it or you're taking a 25-30% hit.

We tell our ie6 users one time (at registration) that their browser is not up to date and that they should upgrade to ff or ie7 to take advantage of all of our features. I've seen this solution that looks nice and unobtrusive: http://www.pushuptheweb.com/

I asked my girlfriend a few months ago why she never upgraded ANYTHING on her computer, ie6 included, and she said she only wanted to check her email and buy things, and that installing updates always took more time than she wanted to spend in the first place, and that she wasn't sure which updates to trust. Many people can't upgrade their browsers, many people don't care to, many people don't know why they should.

oh well. that's the world of warcraft that you play :)

Yahoo uses the Graded Browser Support policy (http://developer.yahoo.com/yui/articles/gbs/) for a reason. It's because we check what % of our users use each browser and support the appropriate ones.

You can bet that if Yahoo is still choosing to support IE6 it's because a decent slice of the internet still uses it. That said, IE <6 is on our C-grade list. This means we actively block them from getting JS/CSS and display a "nag" message with upgrade information.

Just treat it like Netscape 4 or Lynx or whatever -- try to display something meaningful, but don't bother about getting all the details to work entirely right.

Exactly. Many people aren't choosing to still use IE6 - the choice is being made for them. I work onsite in a non-profit sometimes, and they are all still on IE6 until the IT dept decides it's safe/worth it to upgrade. (EDIT: And I do my part to convince them it is safe and worth it.)

I feel like every month or so, a big site says "Death to IE6," then people point out that it's not always an individual decision to upgrade, then someone like marijn points out the most reasonable solution. Make sure the content comes through, make sure it doesn't look completely ridiculous, and educate clients as you go. And if you're lucky, like 37signals, you can decide for your clients. We'll get there eventually.

Bad advice, look at your stats first and then decide. If IE6 represents less than 5% (or whatever your threshold) then treat them like Nutscrape 4. However if it's the majority of your visitors it's not a great idea to piss them all off.

i agree - it's important to remember that large corporations are always behind and for good reason

I am one of two front-end engineers at a major site with 10 million members. We're having this discussion internally. From our side we want to stop because 80% of our time is actually spent on making things work for IE6. It's a huge number. 80% of my salary is being spent on less than 30% of our audience. 78% of our total traffic is IE, with 66% of that being ie7 and the remaining being 34% ie6.

The biggest hurdle we face is that we aren't a small company. We can't make smart decisions like adding upgrade prompts because marketing would throw a fit.

The approach right now we are taking is to get data. Microsoft ends support for IE6 in July 2010 when support for WinXP SP3 ends. After reading about the blogger who found only 2% of his ie6 visits were real, I am now looking into that for us. On our projects, I am keeping track of how much time is spent on supporting it versus ie7, ff, and safari. What it boils down to is that we will only be able to drop support when we can prove it costs us more than we make from ad clickthroughs from ie6 users.

What you you building that requires you spend 80% of your time fixing for IE6? I can't even imagine.

First, 80% is an off the cuff comment/estimate. Obviously its not that high, but it is still very high. I had meant to say 'feels like 80%'.

When you work for a large company and the major decisions on visual design are made by other people, you have little control over how it looks or behaves, and only the freedom to code. Unfortunately, this breeds an atmosphere where people with little concept of the difficulties of implementing cross-browser designs are the ones expecting and demanding that they work in IE6 and "it should look like the picture from the design team."

So yes, a site or feature that takes me hours or a day to code up and be proper in FF, Safari, and IE7 usually adds up to 4x the same amount of time to make it work in IE6. Not always, but often enough that it has become an issue.