How realistic would it be to produce an open source replacement? (Maybe a Kickstarter-funded one, to speed things up, given the fact it concerns every US household?)
You're solving the wrong problem. The issue isn't that tax prep people are making an easy problem seem hard (and are thus ripe for an open source disruption). Intuit (and others) have a vested interested in making it actually hard.
Filing taxes correctly while maximizing benefits is incredibly difficult. There are tons of ambiguous rules and loopholes and fuzzy boundaries and it changes all the time.
It's pretty difficult. E-Filing isn't as simple as a web service.
I believe that if it were to actually do the e-filing, the people who ran the e-filing would have to submit to federal background and clearance checks at those people's expense and the servers it ran on would have to probably get FISMA High certification, provided it actually posted to the Efiling API.
That's just my speculation, fwiw. But I think that's true.
Uggh, from what I've heard IRS makes time management look like a simpleton banging toy trains and making "wuwu" sounds.
And time management is notoriously hard. Not to mention this will need security up the bum.
How much would it cost just to read all the documentation and have the program certified, so it can do what it's supposed to do without writers getting sued?
Might as well at least mention this year's revelation, which is that they fund fake Grassroots campaigns against tax reform: http://www.propublica.org/article/turbotax-maker-linked-to-g...