home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!digex.com!kfl
- From: kfl@access.digex.com (Keith F. Lynch)
- Newsgroups: rec.puzzles
- Subject: "Vesterman's paradox"
- Date: 28 Dec 1992 19:07:12 GMT
- Organization: Express Access Public Access UNIX, Greenbelt, Maryland USA
- Lines: 35
- Distribution: world
- Message-ID: <1hnj90INN4f5@mirror.digex.com>
- References: <92351.132601RVESTERM@vma.cc.nd.edu> <1gp8ttINNik@gap.caltech.edu> <92352.045047RVESTERM@vma.cc.nd.edu>
- NNTP-Posting-Host: access.digex.com
-
- Bob Vesterman, <RVESTERM@vma.cc.nd.edu> claims that it's a paradox that:
-
- 1) "The smallest positive integer too small to be specified in 150 ASCII
- characters" doesn't exist, since if it did, it has just been specified
- in fewer than 150 characters.
-
- 2) There are a finite number of ASCII character strings with length 150
- or less.
-
- 3) There are an infinite number of positive integers.
-
- The resolution lies in the fact that there's no unique mapping of
- character strings onto integers. "The smallest..." only makes sense
- in terms of some mapping. One mapping would be to use nothing but the
- digits 0 through 9, in the usual way. In that case, 10^151 is the
- smallest positive integer which cannot be specified in 150 characters.
- If we use base 256 notation, using the full eight-bit ASCII set, then
- 256^150 is the answer. These two mappings have the virtue of being able
- to represent any positive integer less than this limit.
-
- What if our mapping includes the digits 0 through 9 and the symbols
- + - * / ^ to be used in the usual way? Then the highest number that
- can be represented is 9^9^9^9^9...9^9^99 (note that the number one less
- than this cannot be represented).
-
- Since there are no restrictions on possible mappings, there is no highest
- integer that can be represented in 150 characters. Or in 1 character.
- I could take hundreds of megabytes describing what number I mean by "N",
- and then giving the string "N", representing whatever mind-bogglingly
- large number I defined it as.
-
- Of course you can devise meta-mappings, i.e. languages for describing
- mappings, but that doesn't change the essense of this argument.
-
- Keith Lynch, kfl@access.digex.com
-