Re: gsl_histogram_sum proposal

[Achim Gaedke <Achim dot Gaedke at uni-koeln dot de>]

Brian Gough schrieb:
> 
> The "stat" functions should operate on the histogram as a whole in my
> opinion.

That is a good reason - on the other hand the pdf functions have quite similar
mechansims. The cumulative probability aspect - this is a scientific argument -
is hit here and its my initention for implementing slices of the total sum.
Maybe we will leave it as it is: without slices. Here is the diff for Geralds
proposal, so it could be applied or not.

> In general the user should not need to deal with bin
> numbers, they are not very meaningful.

The discrete aspect can't be denied. Index numbers must be.

> We seem to be missing stat2d.c (I noticed you put the corresponding
> routine in maxval_2d.c).  Would you like to do a stat2d file to
> compute x-y means, sigmas and correlation coefficients?
> 
> --
> Brian

I can schedule this for next week. Are there hints/ideas for numeric
implementation (references, literature)?

Attachment: diff.histogram2
Description: Binary data