Does modeling provide an ROI and how?
I start from the premise that the quality of the outcome of technical processes related to the Test Domain (e.g., integration, verification as described by the technical processes of ISO/IEC/IEEE 15288) is dependent on the quality of the skills of the resource performing the role responsible for the process.
My interest is in the “cognitive processes”, as described by “Bloom’s Taxonomy”, required by a “Role” in the execution of a “responsibility”. “Roles” produce “Outcomes” by performing “Activities”. “Activities” require “cognitive processes” to perform. These “Activities” are the “responsibility” of a “Role”. The quality of an “Outcome” is dependent on “cognitive process” proficiency of the individual performing a “Role”.
So the most important elements of the taxonomy are: role, activity, cognitive process(es) required to support a role activity
outcome * cognitive_processquality = outcomequality
The premise is that models and modeling languages are cognitive process tools. Being proven true, it establishes that there is an enabling relationship between cognitive processes and models and modeling languages. They enable and enhance, thereby creating value.
By establishing an understanding of this fundamental relationship, the value of modeling becomes apparent. It is more likely a better indicator than other metrics of modeling benefits that are being requested (e.g., cost, schedule, quality measures) by much of the management infrastructure.
Models and modeling languages are tools that directly influence outcomes of cognitive processes. By enhancing core cognitive processes, program performance is improved. The relationship between modeling and program performance is not direct, it is a consequence of improving cognitive process quality.
Hopefully, this brings the focus back on engineering fundamentals and why we cannot ignore them. “Fools with tools are still fools”. Modeling is not a silver bullet, it is a multiplier. Zero multiplied by any very large number is still zero.
The individual fulfilling a role must possess the essential cognitive process capabilities demanded by the role’s responsibilities. Modeling enhances cognitive processes, it cannot proxy for them.
Cognitive_Process_Proficiency * Modeling_Proficiency
is a “Leading Indicator” and we can use this understanding to forecast program risk.