In the rather artificial world of Systems Engineering there are a couple of interesting types of inhabitants: Gorillas and Guerillas. Gorillas are the BIG guys: the US Government and the “uber” integrators, the primes. These are the organizations and their members who produce or cause to be produced large complex and / or complicated systems. They use the standard system engineering approach described by the â€œVâ€ (or equivalent); a standard set of artifacts (requirements, architectures, interface control documents, and so on); and a limited set of relatively expensive tools to create and configuration manage these artifacts. On a typical program hundreds (if not thousands) of people work on systems engineering and the tools include requirements management, architecture development and management, custom data bases, computer modeling, and computer aided design to mention the basics. The processes used are invariably proprietary but are always certified, as in CMMI Level n always involve large number of approval boards, and always involve large scale program reviews. In the Department of Defense there are standardized artifacts and reviews that must be adhered to. Some of the authors have sufficient experience in this arena to offer some observations:
- Changes to requirements cause considerable disruption;
- Process often interferes with results;
- More time is spent arguing about process and tools than about results;
- Linking architecture and requirements is difficult if they are built and managed with separate tools;
- Too many boards spoil the product(s);
- Nobody understands how software and hardware are integrated;
- There are at least two sides to every interface and they never agree on the documentation;
- No one knows how to define what constitutes a design;
- Geographical and Organizational diversity make integration really, really hard;
- Systems engineering should not last into testing, must not last into production, and preferably not too far into development.
Guerilla systems engineering tries to deal with the reality of the jungle full of Gorillas and accomplish something useful. It focuses on an aspect of the process that is failing and tries to move outside of the process using small teams of experts to make progress. Some of the authors have also had sufficient experience to offer some observations on this:
- Guerillas are born of desperation;
- Guerillas will only be tolerated for a limited time with a very limited focus;
- Guerillas have very limited resources;
- One case with limited success in the area of design is known;
- Too much visibility will cause their early demise;
- Operating outside of the process will cause their early demise;
- Roadblocks will be thrown up at every opportunity;
- Stealth is essential but impossible;
- Guerillas always loose. HOWEVER, THERE IS HOPE!
The main problem has been the lack of an inexpensive model to allow a small group of organizationally and geographically dispersed experts to keep up with an army of Gorillas. The SPEC Innovations Lifecycle Modeling Language (LML) and tool will empower the Guerillas to win.