Any code authored by or for (exclusively) the government is default open. Usually all it takes is a FOIA request; if you're lucky they're publishing it on github already. However there are carve-outs for areas where making the code public could impede the law-enforcement mission of the entity that uses it, that is, FOIA exemption 7E. Since the IRS knows that if you knew how the audit worked*, then you'd do your taxes "just so" to avoid the thresholds by running the logic yourself, which is not something they'd want to encourage. And it would definitely increase their audit casework load.

There's also the issue that if the code wasn't bespoke but also sold to non-government entities for similar missions (i.e. government does not hold exclusive rights), then it can be protected as the contractors IP. But for the IRS this would be rare, they are pretty unique and often do things their own way.

* You can sort of do this without the code. The IRS is not allowed by legislation to base an audit decision on any information that is not covered by eFile, so contents of forms 1041QFT and 990T, or any attachments to what could have been an electronically submitted form, is out of scope. As long as what you submit in the core set of forms aren't statistical outliers, then you're good.

Generally no. Data you get from FOIA requests is generally limited with what you can do with it. State specific laws, your use-case not withstanding

The IRS is a US federal agency, though. They can claim no copyright on the code. It should be be public domain.

Unless it was written by a contractor who then gave the copyright to the IRS. This is a very common situation. The federal government is not barred from having copyrights.

That would be a fun legal rabbit hole to descend into.

I don't know about fun, but it would definitely be expensive.

"Fun" from a research perspective. (I've got friends who are IP lawyers and enjoy talking about this stuff.)

FOIA results can definitely go into public domain. That's sort of the point.

> That's sort of the point.

It really isn't - the point is freedom access, not free use. Information acquired this way doesn't magically become public domain, it may (or may not) have other constraints on it.

See e.g. https://www.justice.gov/oip/blog/foia-update-oip-guidance-co...