> I find it relevant that when an interrupt driven method has been
> promoted (over Tilmanns objections) a problem get discussed that
> is directly related to using bit-bang.

...however the "overall timing" detail applies to both methods and
was interesting to learn about.

Of course it all depends on the other tasks that have to be done in
the particular application. In mine, there are no such lengthy ISRs,
for example.