On Thu, 21 Mar 2002 22:20:16 Barbara and/or Bill Hooper wrote: >on 3/20/2002 3:43 PM, Terry Simpson ([EMAIL PROTECTED]) wrote: > >> I use multiples of >> 1, 2 and 5. Thus 10, 20, 50, 100, 200, 500 etc. >> >> These are equal intervals on a logarithmic scale > >But they are NOT equal intervals on a log scale. >... >(I've used natural logarithms,to the base e, but similar results arise using >base 10 logs.) > >There are many good reasons for using the series 1, 2, 5, 10, etc. but exact >equality of logaritmic intervals is not one of them. > >Interestingly, the series 1, 2, 4, 8, 16, etc. DOES have exactly equal >logartimic intervals, but this is a series that Terry (and I) depricate for >other reasons. (The interval, in natural logs, is 0.693 for each and every >step.) >... Just a couple of minor observations here. Firtsly, logarithmic scales are usually only employed when one studies phenomena that are exponential in nature. The main advantage of which would be to generate straight lines as opposed to "curved" ones. We really don't care one bit if intervals *in that scale* would be "crooked". What we usually measure is the angle of the line and where it may cross the axis.
Secondly, obviously whether intervals are equal or not will largely depend on the *base* of the logarithm. And here one uses overwhelmingly three bases, ten, e and (with much less application) two. Marcus Is your boss reading your email? ....Probably Keep your messages private by using Lycos Mail. Sign up today at http://mail.lycos.com
