Friday, December 28, 2012

Matlab Wizard


it's all about abstol, she said
noting that 1e-6 was not small enough
for molecules bouncing round and
round off telescope walls.
how obvious I noted as
she lowered this mysterious
parameter to 1e-16, how quickly
then did the problem disappear
as negative values rose
in  positive glee to join their
brethren on the log-log plot.

1 comment:

Jim said...

I do loves me my log-log plots. Scales become manageable and lines become straighter. Much goodness that leads to logarithmic means to quell the outlying data scatter.