About the Development of StatistX

 StatistX
StatistX is written in Objective-C. Objective-C is available on most systems where C is available. The first step was to wrap statist into Objective-C classes that allow for convenient access to statist's functionality. The outcome of this work are the SXKernel-class and the classes Column, DataSet and a couple of helper classes. The SXKernel (only one instance per program) works more or less like a plug-in. It provides the statistical functions as methods and it runs relatively independent. This implementation is solely based on the Foundation package and should therefore run 'almost everywhere'.
The second step was to design a user interface. I decided to develop an interface for Linux, since OpenStep (or whatever Apple makes of it) is not as widely distributed as Linux. I'm using the GToolKit for the GUI. GToolKit wraps the gtk-library into Objective-C classes, based again on the NSFoundation class hierarchy. Gtk is available on Linux and other Unix systems and will work on Win32 in the future. If anybody wants to develop an interface for MacOS X - go ahead! The kernel is ready.
The GUI classes use the GtkSheet-widget to display the data in a spreadsheet like table. The user can select columns and perform statistical tests and regressions on the selected columns. Most of the output is displayed in a simple text-window, since statist produces just text and I would have a hard time re-writing all the output- and result stuff of Statist. However, if available, statist uses Gnuplot for a large part of its output.
See Structure of Major StatistX Classes.

To Do

Well, certainly there is much to do!
  1. Internationalization: Actual internationalization is not supported at all. However, using the Foundation classes it should be possible to change that!
  2. Stability: Ok. StatistX is probably not 100% stable, but it should be.
  3. DnD: There is no drag-and-drop feature. Who has experience with gtk's dnd possibilities?
  4. Documentation: It is now possible to write on-line help. However, it has to be written.
  5. File formats: StatistX reads ASCII files only. It is my plan to implement an HTML-parser that extracts HTML tables and imports them into StatistX. Anything else should be converted into HTML before it is loaded into StatistX.
  6. Macros: StatistX should allow for macro use in future. It is an open question which language should be used for that purpose. I'm open for suggestions.
 StatistX homepage
Screenshot

Download
Structure

Copyright: Andreas Beyer, 1999-2001.   Internal | Infoservice last modified: May 2001

E-Mail: Andreas.Beyer@usf.Uni-Osnabrueck.DE
Access: 2330