[seqfan] Re: Chewing an old bone
N. J. A. Sloane
njas at research.att.com
Tue Oct 25 23:03:06 CEST 2011
Dear Seqfans, Sorry for the delay in responding
to this discussion of plots, but I was in Brazil.
As I see it, the most important points are:
1. One definitely wants to see a log plot if the numbers are large,
2. log(|a|+1) loses the sign if there are negative numbers,
3. asinh(x) does a good job of mimicing log_e(x) for large x,
up to a factor of 2, and
4. No one in the world knows what qlog(x) or quasilog(x) means.
So after extensive discussions, David Applegate and I propose
to do the following:
For sequences with nonnegative terms, the plots will be unchanged
For sequence with large negative terms, there will
be a third plot showing asinh(a(n)). [This will
be on a different scale from the log plot, but the
asinh function is simple and elegant - thanks to
David Wilson for proposing it].
Neil
More information about the SeqFan
mailing list