The thing about supporting legacy IE is to NOT waste time making it perfect, just make it functional. That's what progressive enhancement leading to graceful degradation is FOR. It's not hard to make a page that gracefully degrades functional to legacy IE, unless you waste time doing stuff that has no business on websites in the first place or wastes bandwidth while not adding anything of real value to the site.
Making it perfect in anything less than IE10? Waste of time. Making it WORK in those browsers? That's why we're supposed to be developing with semantic markup, separation of presentation from content, and enhancing functionality with js instead of supplanting it. Hence that pesky "content first!" attitude I'm always on about.
JUST as a properly written page on a normal website if done PROPERLY should be functional and usable without CSS or presentation in the markup - again, the whole "if you are choosing your tags based on what they look like instead of what they mean, you're choosing the wrong tags for the wrong reasons".
Also, don't let people use percentages to lie to you. I love how most people are saying 4% for IE isn't worth worrying about, when the numbers from different sites range anywhere still from 1% to a whopping 25%; but it's IE, people only want the smallest number because "rah, rah, Microsoft is evil".
But think on this: 1% of WHAT? Well... 2.8 billion Internet users is what. So that's... 28 MILLION potential visitors being neglected. It's easy to dismiss a percentage as a percentage, but not when you ask "of what"
See the outright lies and bunko bull people have about IE usage overall; they constantly tout how much 'share' they've 'lost' in the market... Yeah, sure, they've gone from 95% in early 2004 to somewhere between 24% and 52% depending on who's numbers you used. See how NetApplications weighs their values by country resulting in that 52%, since if you're developing for first world nations you don't give a flying purple fish how many Chrome and FF users there are in China.
But, again, a percentage of what? News flash, the pool size grew from 2004 to 2014, let's use the mythical 95% of Jan 2004 to the average across the different stat sites of 25% as of Jan 2014... Well, how many Internet users were there at the start of 2004? 720 million. How many were there at the start of this year? 2.8 billion -- now I'm no mathematical genius, but I'm pretty sure that 95% of 720 million is ~684 million... and 25% of 2.8 billion is 700 million... meaning that while they 'lost' market share, they gained 16 million new users!
Really if you were to graph it by users instead of percent, you'd realize IE use has been pretty flat, mostly just seeing slow gains. Of course, nobody wants to hear that, so they use percentages to card stack their facts.
Opera see's similar stats -- It's easy to dismiss 2%, it's not easy to dismiss 56 million potential users.
It's why I use the term "percenters" -- people who use percentages to push false conclusions without saying "a percentage of what?" It is one of the best lies to promote an agenda. Don't let people use percentages to lie to you. Web devlopers are guilt of this **** all the time -- people browsing with scripting disabled is only 3%, user agents who don't see CSS or style are only 2%, Opera users are only 3%, Safari users are now less than 4%... until they've alienated so many potential users in justifying their sleazy, broken, ignorant practices, they no longer have an audience.