[frossie]

At that point, why not just ask them to provide the equivalent of Turbo-Tax on the web and put the code on github? I mean if they are going to compete with software providers, why mess around with PDFs?

The advantage of tax software is not that it does the caculations for you, it is that it leads you through identifying what facts are important to take into consideration - eg. "Did you sell a house this year", "did you move for your job", "did you buy an electric vehicle" and so on.

[robzyb]

> At that point, why not just ask them to provide the equivalent of Turbo-Tax on the web and put the code on github?

Because we are interested in only the algorithm, not any implementation.

Software blurs the line between algorithm and implementation.

[frossie]

What algorithms are we talking about? The ones that are printed on the tax form? "Subtract box 3 from box 2 and enter it in box 14"? They aren't secret.

The problem people have with tax forms is not the calculations, it is understanding what has to go in the box in the first place. And, for a non-negligible part of the population, enough English literacy to read the form in the first place. Which is why volunteer tax preparation is a thing:

http://www.aarp.org/money/taxes/info-2006/volunteer_aarp_tax...

[Akarnani]

Seconding @robzyb. If the algorithm is out there then people and orgs can create their own implementations that help specific populations without having to license code from anyone or waiting for TurboTax or one of the majors to build a product around the algorithm that supports the needs of the specific population in question.

[amluto]

Also, TurboTax sucks (just try to use it), and I don't really expect the IRS to write anything better. But I do expect people to write open-source frameworks that can parse IRS formulas, and I expect lots of people to try all kinds of innovative ways to wrap the whole thing in a UI.