Structural Modeling Test Bed... Great Idea

[joseph simpson]

Kevin:

Your idea for a structural modeling test bed is very important.

There needs to be an agreed on set of test cases that people can use to exercise their processes and software.

I believe that the test bed should cover all aspects of structural modeling.  Including;

  Basic Structural Modeling (BSM)

  Structural Integration Modeling (SIM)

  Interpretive Structural Modeling. (ISM)

General test cases can be developed for closed source software and associated processes.

General test cases can be developed for open source software and associated processes.

Open source software can be evaluated in terms of software architecture, functions and interaction.

The system organizing relationship properties should be used to create groups of system tests.

For example, given a system organizing relationship that is:

  Transitive

  Asymmetrical

  Irreflexive

Any given system structuring test approach could be restricted to a strict ordering (total order - only one object at each level) or opened up to a general ordering (both total order and/or partial order are allowed).

There are a number of standard clustering approaches that will need to have standard structuring tests, as well.

An organizing relationship that has a asymmetrical property will usually generate some type of linear or hierarchical system structure.  

An organizing relationship that has a symmetrical property will usually generate some type of network system structure.

An organizing relationship that has a non-symmetrical property may generate a range of system structural types ranging from a network to a sequence.

The system organizing relationship is a key factor in creating and organizing system structure tests.

In my work evaluating the published literature associated with ISM, Automated N-Squared Charts and Design Structure Matrices (DSM), I attempted to verify information published in the literature.  Much of the published literature was presented in a manner that made it very hard, if not impossible, to reproduce the published results.

Around 2010, I prepared and published a standard set of matrix examples for use in Automated N-Squared Charts, ISM and DSM software and process verification and validation.  These standard matrix examples would be used with symmetric type of organizing system structures.  These examples were published to reduce the effort needed to copy and paste test data from journal pages.  These examples also reduce the chance of errors in copying data into any computational system.  Please see:

http://systemsconcept.org/sc_sme_v001/SME.html

http://systemsconcept.org/html/Content/sme_cd_data_v002.zip

Similar test sets can be developed for a complete range of structural modeling tasks.

I look forward to working with you and anyone else that wants create a test infrastructure that supports the reproduction of published test results and furthers the open, public evaluation of system science and engineering processes and methods.

Reproducible results are the engine of progress.

Take care, be good to yourself and have fun,

Joe 

 

 

--

Joe Simpson

“Reasonable people adapt themselves to the world. 

Unreasonable people attempt to adapt the world to themselves. 

All progress, therefore, depends on unreasonable people.”

George Bernard Shaw

[joseph simpson]

The activity associated with the structural modeling project created a Ruby code library (gem) to support general ordering of system elements.  General ordering supports both total ordering and/or partial ordering of system structural elements.  

The current project code is available at:

   https://github.com/jjs0sbw/BinaryMatrix

The open source code on Github is  fully integrated into Travis Continuous Integration (CI) system.

All updates to the code base are automatically tested using nine (9) different implementations of the Ruby Language.  The builtin specification test provide some indication of program function.  See:

https://travis-ci.org/jjs0sbw/BinaryMatrix

The next structural modeling code base will be written in Javascript to test out a Javascript library the supports the html5 canvas element.  This is an exploratory test of the html5 canvas element and the Zebkit.js library.  Should be a fun exercise.  See:

http://www.zebkit.com/

We will discuss some of these activities during the Saturday teleconference.

Have fun,

Joe