> (1) I don't think it would break ABI. The additional ios_base::bin flag would have to fit within the current bitmask though.

That's exactly my intuition and that's exactly what's problematic. Any
already compiled operator<< wouldn't check that bit because
historically, it didn't exist.

Anyhow, from what you're saying, it sounds like the worst that can
happen is that your numbers are printed in the wrong base. However,
the paper needs to make sure that that's accurate, and you need to be
quite confident about it.

> Unfortunately I do not have an implementation but it was clarified that I wouldn't need one

You probably don't need an implementation because it's obviously
implementable; the question is just what happens to existing
implementations when you mess around with the flags they store, and
that needs to be in the paper. I wouldn't be the last person to ask
about it.