1997-10-27 - Re: Orthogonal

Header Data

From: Kent Crispin <kent@bywater.songbird.com>
To: Cypherpunks Distributed Remailer <cypherpunks@ssz.com>
Message Hash: dc21011912e40032ceb5f40b7d8b16a609ed87dbc13ee0146e8f448c6afc2c65
Message ID: <19971026163358.52901@bywater.songbird.com>
Reply To: <199710261616.KAA00300@einstein.ssz.com>
UTC Datetime: 1997-10-27 00:44:05 UTC
Raw Date: Mon, 27 Oct 1997 08:44:05 +0800

Raw message

From: Kent Crispin <kent@bywater.songbird.com>
Date: Mon, 27 Oct 1997 08:44:05 +0800
To: Cypherpunks Distributed Remailer <cypherpunks@ssz.com>
Subject: Re: Orthogonal
In-Reply-To: <199710261616.KAA00300@einstein.ssz.com>
Message-ID: <19971026163358.52901@bywater.songbird.com>
MIME-Version: 1.0
Content-Type: text/plain



On Sun, Oct 26, 1997 at 10:16:03AM -0600, Jim Choate wrote:
> Hi,
> 
> I believe the first definition for orthogonal is not the vector or Cartesian
> definition but rather from geometry.
> 
> Where, for example, polygonal means a closed shape made up from many line
> segments, orthogonal means a closed shape made from line segments at right
> angles to each other. The simplest being the square.
> 
> DeCarte used the concept of ortho- to describe the relationship between the
> axis of his measurement system, hence orthogonal. Strictly speaking
> orthogonal is a misnomer and should be orthometric or 'measurements at right
> angles'.
> 
> I am interested in how orthogonal obtained its variety of other meanings. I
> run across it in linguistics, computer science, philosophy, etc. In most of
> them it means some sort of pure or simple relationship. Unfortunately I
> can't find any sort of description of how it got expanded this way.

Sorry I misread your prior post.

The first context where I am aware of this use of the term orthogonal
is from language design -- it was promoted by Niklaus Wirth and other
purists, with languages like Pascal, Modula, CLU, and so on.  Larry
Wall's "perl" language, with its slogan "there's more than one way to
do it", is a direct revolt against the language purists.

The basic idea is that a computer language should have the minimum
number of constructs necessary to span the intended application.  So
for example, you don't provide hyperbolic trig functions, because the
user can implement them using simpler math functions.  On the other
hand, you do supply commonly used math functions that would otherwise
require iterative algorithms.

I do believe the use of the term this way was inspired by the 
notion of a 'basis' in a vector space -- a set of orthogonal
vectors that span the space, ideally, unit vectors.

-- 
Kent Crispin				"No reason to get excited",
kent@songbird.com			the thief he kindly spoke...
PGP fingerprint:   B1 8B 72 ED 55 21 5E 44  61 F4 58 0F 72 10 65 55
http://songbird.com/kent/pgp_key.html






Thread