a tad more on OR-numbrals
Marc LeBrun
mlb at well.com
Wed Jan 2 19:35:48 CET 2002
>=Jens Voss
> In fact, it is not hard to see that [2] is the only OR-numbral
> prime number.
Very nice argument! This is the kind of development of "numbral theory" I
am hoping to see. Each concrete system possesses a rich (unexplored!)
special analog theory, to say nothing of the general questions about what
systems are isomorphic to eachother or to other known algebraic structures...
> Recall that a number is defined to be _prime_ if whenever it divides
> a product, it must also divide (at least) one of the factors.
Not to quibble, but I'm not completely comfortable just accepting this as
the definition of prime. Is there some sort of "classical" development
that necessarily begins with this?
I don't imagine, for instance, that the ancient Greeks would have chosen it
over the seemingly more natural (generalized) definition as "having no
non-unit proper divisors". (Though of course these mutually imply
eachother for ordinary numbers).
This sort of "splitting" is typical of numbrals, and it's not clear to me
which end of these elephants we should start with in building up analog
theories. Maybe both: my preference is to pick definitions that lead to
richer systems, even if some of the theorems for regular numbers have to be
modified or abandoned. But any guidance or suggestions are welcome.
For example not being a UFD breaks the normal Euclidean algorithm, yet the
concept of a GCD as "largest common divisor" still seems perfectly
natural. We just can't deduce all the same consequences. Similarly, every
XOR numbral can be represented by infinitely many positive sums, but if we
restrict it to "proper addends" we can meaningfully define XOR-partitions, etc.
By the way, the sequence giving the number of OR-partitions actually made
me laugh out loud when I first saw it. It looked so absurd (compared, eg,
to the XOR-partitions) that I was sure my program was whacko! But it in
fact turned out to be based on a deeply serious sequence--in the EIS thanks
to Michael Somos--that's been whipped in a blender like mathematical
margarita mix by Zippy the Pinhead. (Is there an alt.humor.sequences?)
Anyway, thanks for the interesting analysis!
Happy New Year!
More information about the SeqFan
mailing list