The conventional wisdom is if 10% of your users are on a particular browser, you should probably support that browser. The problem, however, is that if you build an HTML5/CSS3 Responsive website from scratch you inevitably end up spending a great deal of time putting in hacks just to get things working somewhat similar in IE 8 (which doesn't support a large number of the new standards coming out now). To give an example, I spent over 5 hours yesterday working on a silly image overlay issue because of the weird way that IE 8 handles z-index. If you have the resources, of course it makes sense to build in some type of support for such a large number of users, but if you at a small startup where there is very little time for anything shouldn't you focus on the 90%? I know BaseCamp is doing this with their latest version, but is it crazy to force the use of alternative browsers? Is the huge effort needed to make things easy for that 10% worth it?
I think it all comes down to personal preferences and no you never HAVE TO make stuff work for older, broken browsers. It is perfectly legitimate to ask users to use a different browsers if they are on an old browser.<p>The best way to decide wether to spend time debugging or just screw it would to be ask yourself a couple of questions.<p>How much of the user base will be using old browsers? Think if your user base is mostly geeks / people at home or non-technichal / people using school or work computers (with locked-down policies for downloading software)<p>Can you spend less time on older browsers and make it gracefully degrade without losing functionality? If your website looks less beatiful on an older browser but the functionality still works, you can put a warning stating that the website will look better if the user changes browser, but that no functionality is changed.<p>Can you postpone this debugging to later? Maybe you should focus on launching your product first and once you're mostly satisfied with what you have, your spend time debugging old browsers. This way, you can release quickly for the majority of people that would use your service.
If you don't want the users, ask them to switch browsers. If you want mainstream computer users then support IE8 because it isn't going away any time soon.<p>In my opinion, users will probably lead to more revenue than whatever you are doing with z-index.
Keep in mind that Windows XP users can't upgrade to IE9, they're stuck on IE8 or an alternative.<p>There are still a lot of Windows XP machines out there.
What it comes down to:<p>Is the cost of Supporting IE 8 greater then the revenue an extra X% of users will generate.<p>Obviously this comes down to:<p>1) How much will it cost to support older browsers.<p>2) How many users do you have/what percentage use older browsers<p>3) Actual and expected monetization per user<p>Also bear in mind that users of different browsers may monetize at different rates.
If you do decide to support it, I have found these helpful
<a href="http://css3pie.com/" rel="nofollow">http://css3pie.com/</a>
<a href="http://code.google.com/p/explorercanvas/" rel="nofollow">http://code.google.com/p/explorercanvas/</a>
Do you have any analytics? Standard demographics are ~20-30% IE>9 users.<p>If those users aren't in your target market, I wouldn't worry, but if they are I wouldn't ignore them.