Re: New book: Statistics - A Maxima Companion

Wolfgang, thanks a lot for your work on the statistics companion for
Maxima. I have added it to the list of publications:

https://maxima.sourceforge.io/documentation.html

Very solid work, comprehensive and clear. I have a few minor comments
which I hope you might find useful.

I guess I would recommend using existing functions to the extent that
it's not too clumsy to work them into the examples. There are pdf,
cdf, etc. for a number of distributions in the distrib package.

Also, some of the helper functions could be replaced by built-in
functions. I noticed listmake(x) which is the same as makelist(0, x).
Also, 'ranks' and 'ordering' are now built-in functions, although it
appears 'ranks' is the same as the function 'correspond' and not the
same as 'rank' in the examples.

I think one of the major advantages of Maxima with respect to
statistics is the ability to easily mix numerical and symbolic
computing. I've encountered problems in which that proves to be useful
in cases where the values involved might be either numerical or
symbolic. E.g. a data set comprising [ 11, 5, 13, 17, 8, 5 + y, 2, 7
], or a data set [ a[1], a[2], a[3], a[4], a[5] ], then computing some
kind of general result, and then studying that expression as a
function of the free parameters.

As a part of that, I would recommend applying float or numer less
often; I think the symbolic and exact results would be clearer without
it.

There is, I believe, a lot that can be said about Maxima as a tool for
general data processing, especially of a non-numeric or semi-numeric
flavor. But that's an entire separate book, I suspect, which I might
write someday.

Keep up the good work! It is very valuable to the Maxima community,
and I'm grateful for your contribution.

All the best,

Robert