Robin wrote

> IF A NAND B is supposed to be NOT (A AND B), shouldn't the last line be:
> 
> A NAND B         11111111 11111111 11111111 11011010 <-- correct answer

You are correct! I knew I'd screw it up...

> If that is the correct definition of NAND, then I agree it should be
> changed. However, in that case I would like an alternative way to do what is
> currently provided by NAND: to turn selected bits off with a single
> operation (having no effect if they're already off) - just as OR provides a
> way to turn bits on (having no effect if they're already on). ANDN?

That is an interesting question.  I was approaching this from a digital
logic perspective as I understand the NAND and NOR definition.

Best,
Bill