Re: Has Gnumeric considered wiritng an SPSS clone?

[Jim Robison-Cox <jimrc montana com>]

Hello Uri,

In jumping in here, I should say that I am a statistician, and not a big
fan of SPSS.  Gnumeric is great, in my estimation.
  Several projects to note:
R is a freeware statistics package, which is now as powerful as any
package i know of.  Because it is freeware, hundreds of stat people have
been contributing to it, and it has packages which handle all sorts of
difficult statistical analyses.  See:
  http://www.r-project.org/  or http://cran.r-project.org/

For the occasional user, R has the problem that it is command-line-driven,
no Graphical User Interface.  But several projects are underway to build
one.  See
http://www.sciviews.org/_rgui/
  One project is underway to interface R with gnumeric:
http://www.omegahat.org/RGnumeric/index.html
  but it is not intended (primarily, at least) to supply the type of GUI
you'd like.  Rather it is designed to allow one to register R functions in
gnumeric so that a competent R programmer could use the stat package to do
some computations instead of using the gnumeric engine.

  Several others of the many links on the RGui page would be closer to
your desire, esp RKWard and ObveRsive, both are just getting started.

On Thu, 12 Dec 2002, Uri David Akavia wrote:
> I am a Med student, and we do some work on SPSS (for windows). I believe
> that Gnumeric has a chance of being a clone to SPPS, with several
> (minor) additions. This is because its statistics capability seem above
> those of excel, and close to the capabilities of SPSS.

  One final opinion:  any decent GUI for a stat package should allow
(require?) the user to save the program calls which the GUI generated.
That way you can recall -- a year later-- what analysis you ran.

Cheers,
Jim Robison-Cox