
Re: R > oo
Posted:
Mar 2, 2013 10:28 PM


On Mar 3, 1:05 pm, "Ross A. Finlayson" <ross.finlay...@gmail.com> wrote: > On Mar 2, 3:01 pm, Graham Cooper <grahamcoop...@gmail.com> wrote: > > > > > > > > > > > On Mar 3, 7:35 am, "Ross A. Finlayson" <ross.finlay...@gmail.com> > > wrote: > > > > On Mar 2, 12:30 pm, Graham Cooper <grahamcoop...@gmail.com> wrote: > > > > > On Mar 2, 7:28 pm, Rupert <rupertmccal...@yahoo.com> wrote: > > > > > > On Friday, March 1, 2013 9:39:31 PM UTC+1, Graham Cooper wrote: > > > > > > A LIST oo ROWS LONG! > > > > > > > 0.00... > > > > > > > 0.00... > > > > > > This post is incoherent dribble. > > > > > I showed a partial infinite list of reals. > > > > > 0.00.. > > > > 0;00.. > > > > .. > > > > > Do you have an ANTIDIAGONAL function to support your claims it is > > > > incomplete? > > > > > What is your antidiagonal function? > > > > > Herc > > > > How would you establish that the expansions you begin to detail would > > > map on to any segment of R? > > > > Well, that gets into whether the function, that makes a list these > > > expansions, has as a range, an interval of reals. > > > > So, look at the equivalency function, as I call it, it's quite well > > > defined, it goes to one, and in binary there's only one antidiagonal, > > > and it's one. > > > > I'll agree that a more carefully defined function, that would have as > > > each initial segment of each initial segment, of a matrix of values of > > > the expansions, zeroes, with the only antidiagonal in binary being > > > ones, with real value one, may go from zero, to one. > > > > And: only one does. > > > > Then, for a conscientious mathematician, formalist yearround, that's > > > compelling. > > > > There are lots who would work in foundations, but transfinite > > > cardinals aren't used in real analysis, or continuum analysis for > > > applications and physics. And, physics needs new methods to explain > > > results of experiment. And, results in the digital are available via > > > asymptotics. Good day. > > > > Regards, > > > > Ross Finlayson > > > So your argument is: > > >  > > > Given this sample of a list oo rows long IN BINARY > > > 0.00.. > > 0.00.. > > .. > > > We know 0.11.. > > is missing from the List by Extrapolation over all digit positions. > > >  > > > Given this sample of a list oo rows long IN TERNARY > > > 0.00.. > > 0.00.. > > .. > > > We know > > 0.11.. > > 0.12.. > > 0.21.. > > 0.22.. > > > are missing from the List by Extrapolation. > > >  > > > Given this sample of a list oo rows long IN BASE 4 > > > 0.00.. > > 0.00.. > > .. > > > We know > > 0.11.. > > 0.12.. > > 0.13.. > > 0.21.. > > 0.22.. > > 0.23.. > > 0.31.. > > 0.32.. > > 0.33... > > > are missing from the List by Extrapolation. > > >  > > > Since all above arguments must hold, the latter more absurd ones are > > enough to throw doubt on Cantors Method  which is actually just > > induction over ALL sizes of FINITE lists, since no_new_digit_string is > > calculated in the AntiDiagonal on some infinite lists of reals. > > > Herc > > www.BLoCKPROLOG.com > > Formal rigor is a goal of foundations. > > Here then a consideration of the information content in various bases > or radices used for expansions (of fixed magnitude precision) has that > while any finite base can be used to represent any given (standard) > real number between zero and one in an infinite sequences of elements > of that base, it takes more information to store elements in a higher > base. In this sense binary gets the most elements of the expansion > for a given range of precision. For example with twenty bits in > binary it takes only 2^20 distinguished elements to comprise it, for > twenty bits in decimal, 10^20 which is already 5^20 times as great as > 2^20. Thus as there is approximated a real value to a given number of > digits, that precision is reached first in binary. Then in a sense, > writing the digits in binary is the most efficient way to represent > the number. > > Then for this equivalency function, representing the numbers in > ternary, quaternary, and so on, bases > 2, has that the antidiagonal > could be at minimum .111_b, or sum_i=1^oo b^i, the sum of negative > powers of b, or 1/(b1). Then as the base goes from 2 onward, the > interval would range to at most 1/2, 1/3, .... Yet, with an infinite > radix, it would go back to being one. > > This loophole, this special case, this prototype of functions defining > intervals as arrayed lines of points in drawing them as arrayed from > point to point, has simply that Cantor's results on reals are that > they are only arrayed this way, countably. These are for the number > theoretic considerations, for the settheoretic considerations there > are as well notions of what the real numbers are as extra "complete" > ordered field, or what the linear continuum is, besides as of the > "wellfounded" infinity. > > So, for a function that has each initial segment of an expansion, for > each initial segment of expansions, having at each modulus zero, and > that the only symbol in the language for all bases not zero is one, > the range of this particular function goes to one, which is greater > than or equal to the value everywhere different of each of the initial > segment of each of the initial segment, and simply scaling the range > (of EF) to the reciprocal of the base sees the antidiagonal: nowhere > before the end of the list, which it is. The limit IS the sum. > > For formalists, "exists" is particular. > > Regards, > > Ross Finlayson
You are still left with the Absurdity that given only the following information
0.00... 0.00.. ..
in Decimal, 80 different reals are calculated USING_CANTORS_METHOD that are "missing" from that List.
The Method applied is "anythingbutthatdigit" or NOT(DIGIT) which does look more rigorous in Binary, but you don't get to select which base a general proof on reals looks ok in.

I am expanding logic into Natural Language and the central reasoning heuristic seems to be
NOT ALL == SOME DON'T
!(ALL(X) p(X)) <> EXIST(X) ~p(X)
I already programmed Prolog to solve that Not All Natural Numbers are Even by finding a Natural Odd number s(0), a counterexample.
e(A, not(evens)) < e(A, odds)
I'm now looking at : NOT(EVERYTIME(X)) event(...) <> SOMETIMES( Opposite(event)) (NOT(EASY) method) <> SOMETIMES(hard, method)
Also TOO(BIG) seems to work like NOT(BIG)
and TOO(LATE) et al in Temporal Logic.

I mention this as not can work with discrete sets
color( C, not(red) ) < color( C, blue) color( C, not(red) ) < color( C, green)
where NOT(COLORS) have multiple possible values.
this is like moving to a higher base,
although
digit( D , not(0) ) < digit( D, 1 )
and ANTINUMBERS is not likely to help the dismal array of R>oo believers...

the main argument will always be that a SET of REALS has infinite possible paths
from digit1 to digit1 to ....
the scope of anti(digits) to select one from oo many choices is far greater than the base.
Herc  www.BLoCKPROLOG.com

