We had an interesting session with a customer the other day. The discussion centered on how Cognitive Systems Engineering ‘connects’ with the DoD Architecture Framework (DoDAF). After some interesting moments of talking past each other, the discussion settled on two main points:
- In order to be relevant in major DoD systems programs, Cognitive Systems Engineering will have to find a way to integrate itself into the DoDAF mindset,
- The Applied Cognitive Systems Engineering -ACSE(tm)- methodology (as an instantiation of Cognitive Systems Engineering) delivers revolutionary, high quality designs without the heavy overhead burden of the DoDAF processes.
ACSE was expressly developed to ‘fill in the gaps’ in IEEE Systems/SW Engineering processes. In essence, it was developed as a COMPLEMENT to such processes, not as a replacement. Internally, for relatively small scale systems, we find it produces powerful, revolutionary Decision Support Systems often ready for SW implementation. For some systems, ACSE gets augmented with other IEEE processes and artifacts (i.e. Interface Specifications (for data exchange APIs) and Software Requirement Specifications (SRSs) ). ACSE has also been worked alongside Rational’s Unified Process (RUP) with similar successful results. In short, ACSE delivers the powerful, revolutionary advances in Decision Support System capabilities that DoDAF, RUP, Agile, Extreme, etc. all lack, and depends on a quality development process to ‘deliver the goods’ in an actual working software system.
Many organizations adopt a process like DODAF and quickly lose sight of the objective. In such ‘artifact heavy’ frameworks, the artifacts themselves can become the focus, rather than the quality of the support being provided for the user/for mission success. The several times we’ve been part of a DODAF based program, the form and compliance of the artifacts overwhelmed any attempt to ensure adequate fundamental decision support was built into the system, each time resulting in a brittle Joint Cognitive System. In essence, these were cases of “building the wrong things well.”
It takes a pretty agile mind to make this all work out. Superficial understanding of DODAF fails to understand the inherit brittleness of its underlying Waterfall Methodology roots. Blind execution of RUP similarly loses sight of its spiral/object oriented biases. Taking a step back from these (and others) the broader context comes into view: these are essentially ‘construction’ methodologies that profit greatly when complemented by a generative process for discovering the key functional capabilities that are essential for the user to succeed with their work in the real world. Without such a complementary, user-centered process, ‘weak methods’ fill the void: RAD/JAD, XP, Requirements Decomposition, etc. But in order to integrate Cognitive Systems Engineering some ‘bending’ of the artifacts and task sequencing is required. As a simple example, the unique capabilities of proper support for Directed Attention has never appeared in any of the OV-x documents of a DODAF program. (Here, I hope we’re wrong and someone will offer one up.)
In summary, there is really no argument at all, it’s a case of ‘violent agreement.’ Cognitive Systems Engineering (in our case instantiated by ACSE) provides the breakthroughs in the necessary Decision Support to improve the robustness and mission effectiveness of the Joint Cognitive System. Systems Engineering processes are essential to constructing the computer (HW and SW components) of that JCS, and become increasingly critical as the scale and scope of those components becomes large. All it takes is to ‘allow’ the CSE insights to ‘enhance’ the various artifacts with good Decision Support System capabilities. Ultimately it’s about “Building the right things, well”.
Stumble it!