jos hendriks wrote:
Hallo Morgan
Thank you for the clear reaction Sorry I missed your factor 1/2.
I suffer from an embarrassing psychosomatic condition called "premature optimisation" ... that's why I showed an interest in Gerald K. Dobiasovsky's technique (A fellow sufferer!).
I prefer to use names like mu and sigma, because they are widely agreed upon.
So do I; I only changed "mu" to "o" in the gaussian distribution because there the two are equivalent, but the same isn't true in general. The main reason is because the effects of 'mu', 'zeta' and 'sigma' are already familiar to those familiar with the distribution in question. Maybe I should take your suggestion on - be even more relaxed than I already am and allow them to become complex... There are always all sorts of directions one can take, and one can take only a finite number of them ...
I took a look to the log normal distribution and found it interesting. Maybe you appreciate this:
See? Told ya! Midgets! I _wasn't_ imagining them! :) Had a look around the image; saw, among other things, a vive voce in progress. Didn't record that one, though; recorded this one instead: Taken_in_Lieu { ; A Critical Point in Negotiations ; ('snot your job) ; After Jos Hendriks, 2002 ; Version 2001 Patchlevel 8 reset=2001 type=formula formulafile=distro.frm formulaname=lognormal_distm passes=1 center-mag=0.1353/0/1.5 params=-1/-1/0/0/0/0/0/0 float=y outside=fmod colors=000_XR<43>_hj_hj_hk<2>_il_il`jk<60>tnYunYunX<2>vnXvnXumW<73>ZZJYZ\ JYZJ<3>XYIFN9WYI<52>FN9 } Seems to give the engine a bit of grief here and there - as tends to happen when dividing by (really small) z. Morgan L. Owens "_'Allow'?_"