When to Begin Modeling?
With the joys of 2020, idle time perusing the internet has been in abundant supply. Aside from the excitement watching your standard cat videos, random characters' live-streams, and Tik-Tok dancing competitions, I've had ample time to take advantage of watching/reading posts and presentations on where modeling fits into a project's lifecycle.
Several modeling SMEs tend to argue get started modeling as soon as possible. Model as much as necessary, and even if the variants of the systems architected are unsuccessful at meeting project/programmatic requirements, retain them because they could be used later. Supporting this school of thought, I developed and published the Model of Models Methodology for Model Based Systems Engineering which has been presented to, and adopted by, various circles of modeling experts to accomplish the goal of enabling reuse of your model data to reduce re-work and reduce the amount of time required to architect a system.
However, from researching more agile methodologies in the software and business spaces, I must question my previous mind-set and admit that even in my short career as a modeling champion, there have been times where too much time has been spent modeling details of a concept which was built on sand. I must admit, if the concept has been firm, modeling has proven to be quite successful in enabling analysis, consensus, and preventing errors early on. However, when experimenting/inventing, I can't claim to have been as successful in my modeling endeavors as while the concept is evolving the number of restarts, and amount of re-work in the modeling arena has been significant.
So humbly, I open this to my peers, when is the most appropriate time, in your opinion, to begin modeling (not diagramming, but in-depth modeling) efforts? Relating this question back to standard acquisition terms to kick off discussion, would you find value in modeling before an OPSCON, in parallel with an OPSCON, after an OPSCON, or even later in the project/program lifecycle?
--
4 年I can't really add anything to Daniel's brilliant response, but I can give you my take on OPSCON-related modeling. These is a significant lack of attention paid to higher level and operational modeling in my experience. If you can get access to the subject matter experts, I think that modeling should occur before an OPSCON is written and in parallel with its development. That modeling can serve to inform a better and perhaps more thorough OPSCON and would itself be a great place to start for additional modeling at lower levels.
Systems Engineering, Digital Engineering, MBSE
4 年There are two types of modeling that come to mind when answering your question: divergent modeling and convergent modeling. Early in the development lifecycle of a system or during the establishment of an enterprise or program office, divergent modeling seeks to find as many potential answers as possible and open up the solution space so that needs can be defined, existing technologies can be documented, and trade-off analyses conducted. Once the problem is understood, either through a set of peer-reviewed requirements or with a concrete business or operations plan, convergent modeling will begin. Convergent modeling must end with a final product or solution identified, although the term "convergent" does not preclude additional solutions from being identified when new details about the needs are identified. For divergent modeling, I would begin modeling early and often, with an understanding that not everything should be kept. Many times, early ideas are scrapped after peer-review. For convergent modeling, the model itself does not need to begin until there are a clear and concise set of requirements.