Living in the Past

[Mark Greenfield]

There was just an interesting post on the uwebd list asking the question "is it time to kill IE6" - see http://www.stoplivinginthepast.com/.

Your thoughts? From a pure usability perspective, is it acceptable to not support a specific browser regardless of how small the number of users?

[Curtiss Grymala]

This decision absolutely needs to be based on the percentage of visitors that are using IE6 (and, to a certain extent, who those visitors are).

For us, unfortunately, it is still completely unrealistic to phase out support for IE6, as 25% of our IE visitors still use IE6. IE users make up 78% of our total visitors - doing the math, that makes about 18% of our total Web site visitors still using IE6. That's still a higher percentage than all of our Firefox visitors, and pretty darn close to the total of all non-IE visitors. That's a strong argument for us to continue supporting IE6, no matter how painful it may be.

However, that said, that does not mean you have to spend hours and hours making sure that everything looks perfect in IE6. Just make sure that nothing looks broken, and that everything looks close to the way it really should. If a div element is a pixel or two too far to the left, or a 700 pixel div is only filling 650 pixels of the screen in IE6, I wouldn't fret too much about it.

On two of my hobby sites, however, the incidence of IE6 users is much lower (less than 5% on one site and just under 8% on the other - though, on the first site, the incidence of IE users is right around 25% and on the other site the incidence of IE users is less than 50%), which would make it much more realistic for us to phase out support for IE6.

Also, even if the percentage of users visiting your site with IE6 is extremely low (less than 3%), you still need to make sure that none of the important people in your group are using it. For instance, how embarrassing would it be to stop supporting IE6 only to find out that the president of your institution is still using IE6 for some reason.

[Chas Grundy]

This is similar to the teacher's dilemma of whether to teach to the top or the bottom of the class. It's hard to meet all of the students' needs with the same lesson plan.

When we dropped IE5/Mac support, we had three things going for us: 1) it represented a tiny portion of users and was steadily declining, 2) Microsoft dropped support as well and we could point to the inherent risks, and 3) it cost a ton more to support - sometimes 30% more time. Since we charge for our work, this was an easy decision for clients to understand.

FWIW, IE6 represents a significant portion of our traffic, so we won't be dropping it just yet. But it's also not holding us back in any serious way - we know how to support the browser differences without sacrificing features. Sometimes there's a conditional stylesheet or special Javascript (png fix!)... but IE6 is still in our list to test against.

[Derek Pennycuff]

Officially, my cut off is 2%. But I test in Safari and Opera and those are both well below 2% for us. Most on campus computers are still running IE6. In the redesign I'm working on, IE6 gets a lot of the bells and whistles hidden from it, so that's been a real bummer when I've gone to demo it around campus. We've got a new IT director with a plan to get us all on IE7 next month. I'm anxious to see what that does to our IE6 numbers. Right now we're at 22% by visits. I don't think the on campus change will drop that to 2% or below, but any dip in the number will leave me with less professional/ethical guilt over serving a nerfed experience to those users.

[Skip Knox]

Progressive enhancement. Build sites that work in all browsers and work better in current browsers, and forget about browser versions.

Easier said than done, of course, but I really do feel that the goal of all designers is to get beyond this nonsense about browser version. The goal is device independence, as it has always been.