Xref: utzoo sci.math:3375 sci.physics:3228 sci.electronics:2821 sci.philosophy.tech:610
Path: utzoo!mnetor!uunet!lll-winken!lll-tis!ames!pasteur!agate!eris!doug
From: doug@eris (Doug Merritt)
Newsgroups: sci.math,sci.physics,sci.electronics,sci.philosophy.tech
Subject: Re: distributed transformations (request)
Message-ID: <8859@agate.BERKELEY.EDU>
Date: 18 Apr 88 02:25:24 GMT
References: <8694@cisunx.UUCP>
Sender: usenet@agate.BERKELEY.EDU
Reply-To: doug@eris.UUCP (Doug Merritt)
Organization: University of California, Berkeley
Lines: 23

In article <8694@cisunx.UUCP> vangelde@cisunx.UUCP (Timothy J Van) writes:
>I am interested in finding as many and varied examples as possible
>of functions or transformations with either or both of the following [...]
>
>(1) equidistribution: each output element depends on the whole input

Try Walsh functions (like Fourier but with square waves instead of sine).

>(2) Redundancy: the input information is recoverable not just from the

The theory of error correcting codes applies. What I've read about
it is based on the idea of N-space spheres/circles. Each encoding
represents a point in the space. Each point is surrounded by a circle
of constant radius. If the circles overlap, a bit error is regarded
as a transformation that takes you into the next circle over, and you
can't correct it. If they don't overlap, an error moves you over a
little but it's still unambiguous which circle you're in. Sorry
for being vague (and any errors), I'm not an expert on this.

Laplace transforms in general would probably interest you, too.
How about matrix inversion?

	Doug Merritt		doug@mica.berkeley.edu (ucbvax!mica!doug)
			or	ucbvax!unisoft!certes!doug