An effective approach to drive the challenging tasks of improving Systems Engineering in a structured way is the Systems Development Taxonomy (refer to Figure hereafter). This classification has been established based on many different domains, projects and customers. It determines seven practice stages and also helps in laying out a path for the incremental adoption of Model Driven Systems Engineeing (MDSE) within an organization.
The journey starts at Stage 1 with the simple idea and may finally lead to Stage 7 on the other end of the spectrum with the tested and proven model. The importance and value of each of these consecutive 7 stages is given in order to continuously improve the Systems Engineering capabilities.
The System Development Taxonomy is field agnostic: It can be applied on each system development phase and is applicable to every domain – being it mechanical, software, analogue and digital electronic, optics, etc. Let us briefly go through the stages:
Stage 1 focuses on the idea, stages 2 and 3 on text. These stages are without modeling and are characterized by no or just reduced traceability.
- At Stage 1 there are loosely connected ideas which can be expressed by undemanding tools such as MindMap. Only the idea owner is able to deal with it, which hampers communication.
- Stage 2 is the result of writing such ideas in the form of linear text with text processing tools such as Word, making the idea more explicit, readable and more understandable by other people than their author. Also, it is now possible to manage document versions.
- Stage 3 is characterized by a structured set of text pieces. These pieces can be named requirements. Traceability is now possible among these pieces. Tools to improve traceability and management are e.g. Doors, Doors Next, Polarion, Helix RM etc.
Stages 4 to 7 are with the support of a model. It enhances the capabilities of the engineering process through the whole development workflow.
- The Descriptive Model at stage 4 is the entry point for modelling. In addition to textual, it allows a graphical representation of the system also from a behavioral and architectural point of view. It highlights the interactions between the different elements composing the system. Language specific tools are required such as Visio, MagicDraw, Enterprise Architect, Catia, ProEngi-neer or IBM Rhapsody support this approach. Some of these tools support you in going all the way towards Stage 7, some are limited to Stage 4, and some go some way with add-on tools.
- The Executable Model at Stage 5 allows the complete and executable description of the system in scope. The model of the system has to be executable, i.e. runnable and allowing to perform a system simulation. It enables the user to «see» the system living and can interact with Human Machine Interfaces, virtual mockups, etc.
- The Validated Model at Stage 6 includes a formal and complete description of the tests and allows to verify that the (model of the) system does everything what it is supposed to do.
- The Proven Model at Stage 7 allows to verify that the (model of the) system does not do more than it should, i.e. doing the right things right.
It depends much on the given context which stage your development is and should aspire to be. Important is the movement from left to right, i.e. from Stage 1 towards Stage 7. However, it is imperative that this evolution needs to be taken incrementally and in small steps with a sound vision and management support.
Many companies try to skip certain stages or want to make too large steps and therefore do not reach their goals. Companies taking their present stage into account and taking coordinated steps to reach the next stage have shown improvement by factors, creating as little waste as possible while leveraging value as much as possible.
Stage 4 is the tipping point and opens the strategic opportunity to harvest from MDSE. For many companies modeling is focused on descriptive tasks, i.e. supporting better documentation. The higher potential, however, lies in Stage 5, meaning putting simulation and the executable models in the center of your activity and opening the path to early validation (Stage 6). Validating the model is by far less expensive than using prototypes. Also, model validation allows you to validate very complex scenarios before realizing prototypes. This become s exponentially more valuable as the complexity of your system increases. This Stage 6 is required to reach the Digital Twin and thus allows you to experiment, simulate, test and develop more virtually.