Re: Canonical form with NOR or NAND?

http://groups.google.com/group/sci.logic/tree/browse_frm/thread/a25354b01e1131fc/b9ed37e1a346ec2f?rnum=1&_done=%2Fgroup%2Fsci.logic%2Fbrowse_frm%2Fthread%2Fa25354b01e1131fc%2Fb9ed37e1a346ec2f%3Ftvc%3D1%26#doc_ba1da0109c1d8f4d

So if you start with a pure NOR formula, not already in canonical form,
convert to NOTs and ANDs, convert that to CNF, then use the above rule
to immediately get a 2-level normal form in NOR.

Mitch

Thanks. I'd thought of that. But what motivated my question was
thinking about what representation to use for propositional formulae
in a computer program. If the procedure to conjoin some formula
(representing existing knowledge) with a new forumula (representing a
discovery) and maintain cononical form, is going to be to convert each
whole forumula from NOR form to a normal form with AND, OR, and NOT;
then conjoin the formulae in that form; then finally, convert back to
NOR form again, then I might just as well represent all formulae in
AND, OR, and NOT all the time. Nothing has been gained from adding
NOR form to the design. I'm wondering whether a simpler system could
be realized on the computer using NOR all the time vs. using AND, OR,
and NOT all the time.

.