I still think what I said was correct.

> On the other hand, undecidable has a sharper meaning in computation contexts as well as constructive logics without excluded middle. In these cases we can comprehend the “reachability” of propositions. A proposition is not true or false, but may instead be “constructively true”, “constructively false”, or “undecidable”.

Yes, but that just means that independence/undecidability depend on the proof system, as I said before. So when using a constructive proof system, more statements will turn out to be undecidable/independent of a theory than with a classical one, since the constructive proof system doesn't allow non-constructive proofs, but the classical one does.