From: Vaughan Pratt <pratt@cs.stanford.edu>
To: categories@mta.ca
Subject: Re: Real interval halving
Date: Sat, 01 Jan 2000 14:03:53 -0800 [thread overview]
Message-ID: <200001012203.OAA16605@coraki.Stanford.EDU> (raw)
In-Reply-To: Your message of "Sat, 01 Jan 0100 11:37:15 GMT." <E124MqR-0006i7-00@owl.dpmms.cam.ac.uk>
Thanks, Peter. One of these days I'll learn to stop sending email after
midnight. Continued fractions provide equally good coalgebraic structure
for both our product-with-omega functor (it's one of the examples in
our paper) and Peter F.'s X v X functor. Either midnight madness or
sheer forgetfulness must have possessed me to malign its applicability
to the latter.
While I have that excuse handy let me also repair my description (in
the same message) of halving nonzero reals as right-shifting with sign
extension: following the shift the second bit must then be complemented.
Thus + (1/2) halves to +- (1/2 - 1/4 = 1/4) while +++ (1/2 + 1/4 + 1/8
= 7/8) halves to +-++ (1/2 - 1/4 + 1/8 + 1/16 = 7/16). In the special
case of 1 as ++++... forever, +-+++... equals + (1/2), and dually for -1.
Contorted fractions make an earlier appearance in Conway's On Numbers
and Games (1976) (Winning Ways is 1983). I hadn't realized Norton was
involved there: Conway credits several things to Norton in ONAG but I
guess he must have forgotten that one.
At the risk of turning this thread into a complete tangent space, yet
another construction of the group of reals is as the quotient G/H of the
pointwise-additive group G of bounded integer sequences by the subgroup H
consisting of those sequences b of the form b_0 = a_0, b_{i+1} = a_{i+1}
- 2a_i for some a in G. This definition, which avoids detouring through
the rationals, resulted from my mulling over a talk at MIT by Gian-Carlo
Rota in the early 1970's on representing reals as sequences of bits.
I mentioned it at a recent theory lunch talk and Don Knuth mulled it
over and came up with the idea of modifying the boundedness condition
to allow G to be a ring thus making G/H a field (as an alternative to
taking product to be the unique bilinear operation * satisfying 1*1 =
1), see Problem 10689, American Mathematical Monthly, 105(1998), p.769.
I would love to know whether this construction can exploited in a
coalgebraic setting.
Vaughan Pratt
next prev parent reply other threads:[~2000-01-01 22:03 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2000-01-01 11:37 Dr. P.T. Johnstone
2000-01-01 22:03 ` Vaughan Pratt [this message]
2000-01-02 19:12 ` Michael Barr
2000-01-04 0:35 ` Schanuel's reals Ross Street
2000-01-04 19:44 ` Real interval halving Vaughan Pratt
2000-01-04 20:18 ` Vaughan Pratt
2000-01-01 23:17 ` A couple of Y2K glitches Vaughan Pratt
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=200001012203.OAA16605@coraki.Stanford.EDU \
--to=pratt@cs.stanford.edu \
--cc=categories@mta.ca \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).