Verification and Validation for Prediction Software

Greetings:

I have a strong conviction that the engineering community and the software development community are equally in the dark when it comes to software testing and evaluation (T/E). Nobody knows how to do it, so nobody requires it; since nobody requires it, nobody does it; since nobody does it, nobody knows how to do it; since nobody knows how to do it...[Green and Keys, 1987]. It is of concern that as we get into more and more complex software development and applications (e.g. advanced/autonomous traffic control strategies) that both the current philosophies of software development and software T/E must change or we in the highway business could potentially have our own Cali and Guam crashes (the August 1996 fatal airline crash in Cali, Columbia caused in part by software deficiencies and the August 1997 airline crash in Guam most probably caused by a failure in air traffic control software). New programming techniques and new chip manufacturing--advancements of OOPS, intelligent software and the coming "systems-on-a-chip" [in the next generation of systems chips most circuit nodes will be inaccessible from the outside and will therefore be untestable. Built in self-test will be heavily relied on. If you don't design it right before you build it you are in big trouble.]--will change software development and drastically complicate software T/E. As a very preliminary step in addressing this problem, a draft document on PRODUCING CORRECT SOFTWARE has been prepared. The purpose of this communication is not to present an official document, but to share a work in process and to solicit advice. It is the intention of the Development Team to produce a quality product that will truly be of value to those developing and testing software. Thus I encourage you to critically review this incomplete draft document and provide any suggestions for improvement and expansion. We want to hear the bad news as well as the good so that we can improve this document. It will be impossible to respond directly to every comment, but be assured that every comment will be reviewed and given the consideration it deserves. Thank you in advance for your consideration.

Sincerely yours,

James A. Wentworth
Chief, Advanced Research Team
Office of Safety and Traffic Operations R&D
Federal Highway Administration

[N E X T]