Re: orthogonal specs

Albert Lunde (Albert-Lunde@nwu.edu)
Sat, 19 Oct 1996 01:22:27 -0500 (CDT)


Message-Id: <199610190622.AA257586147@merle.acns.nwu.edu>
Subject: Re: orthogonal specs
To: davidp@earthlink.net (David Perrell)
Date: Sat, 19 Oct 1996 01:22:27 -0500 (CDT)
Cc: www-html@w3.org
In-Reply-To: <199610190419.VAA04805@iceland.it.earthlink.net> from "David Perrell" at Oct 18, 96 09:14:00 pm
From: Albert-Lunde@nwu.edu (Albert Lunde)

> > >From the jargon file (aka the Hacker's dictionary):
> > http://www.ccil.org/jargon/jargon_30.html#TAG1308
> A fabulousity. Thoroughly orthogonal to the American Heritage
> dictionary definition:
> 
> Mathematics.
> Relating to or composed of right angles.

The jargon file quote is based on a much better mathematical definition.

The American Heritage definition is phrased about at the level of 
high school geometry.

The usage that the jargon file quote alludes to extends upwards
into linear algebra, statistics, and just about any abstraction
that supports scalar products and euclidian norms.

The usage in statistics provides a concrete reference, because
independent variables are orthogonal in a particular sense (correlation).

The bit about right angles falls out as special case when you reduce it
to ordinary 3-dimensional geometery.

As mathmatical metaphors go, "orthogonal" in the sense cited in the jargon
file seems fairly well founded, to me.

-- 
    Albert Lunde                      Albert-Lunde@nwu.edu