The question [Is it illegal to rewrite every line of an open source project in a slightly different way, and use it in a closed source project?][1] makes me wonder what would be considered a *clean-room implementation* in the era of open source projects.

Hypothetically, if I were to develop a library which duplicates the *publicly documented* interface of an open-source library, without ever looking at the source code for that library, could that code ever be considered a derivative work?

Obviously it would need the same *class hierarchy* and *method signatures*, so that it could be a drop-in replacement - could that in itself, be enough to provoke a copyright claim?

What about if I used the test suite of the open source project to verify whether my *clean* implementation behaved in the same way as the original library? Would using the test suite be enough to *dirty* my *clean* code?

<sup>As should be expected from a question like this, I am not looking for specific legal advice, but looking to document experiences people may have had with this sort of issue.</sup>

  [1]: https://softwareengineering.stackexchange.com/q/86754/22493