As I was preparing for an upcoming presentation, I stumbled on a graphic that I always thought was one of the best ones in my endless collection of Powerpoint slides. This particular graphic portrays the evolution of engineering modeling software and I always thought it was an incredibly impactful and clear view on a very complex topic. Unfortunately, I really can’t take any credit for it. The basic concept was created by Mr. Alex Ohata of Toyota. I remember the first time I saw it at a conference. It really was one of those light-bulb moments where the Universe unfolded as it should … and now I pay due homage to this work of scientific art.
The graphic tells the story of how the computational side of engineering modeling evolved over the ages – from the early days of the first FORTRAN compilers through to the introduction of numerical libraries, through in state or signal-based simulation tools, to physical modeling tools of today. Click on the diagram to get a close up look.
I think the graphic says it all and that’s the quality of an exceptional technical graphic. For me, it was the perfect balance of
- Temporal dimension (column to column)
- Process dimension (project phases contained in each column)
- Complexity dimension (people you need to work on the project)
- Organizational dimension (the order that the people are listed)
- Solution dimension (human vs. computer effort -- colors)
- Context (examples of software for each grouping)
The Evolution of Engineering Modeling and Simulation
Even a quick look at this graphic immediately tells a rich history of engineering computation. In the early days, it took teams of talented people to run a basic simulation. Why? Because computers were essentially massive programmable calculators, and in order to configure them to solve practical problems you literally had to translate the huge body of human knowledge encapsulated in a typical scientific problem into an airtight digital procedure. Furthermore, the end user – the engineer -- is nominally three degrees of separation away from the computer and the results. He/she will need a modeling expert who then depends on a math expert who will need the input of a numerical expert, just to operate and glean insight from simulations. And for this reason, the vast majority of the tasks are handled by humans and only the final execution of the program is relegated to the machine.
Today, we see a different story. With modern physical modeling tools like MapleSim every task up to and including the development of the model equations can be effectively automated. The computer does pretty much all the work that can be mechanized. What remains on the human effort side is then the stuff of intuition, insight, and experience … things that computers generally stink at.
In a recent seminar, I was asked whether I believed that we had reached a natural limit and indeed intuition and insight can never be automated. My gut says yes but I think that’s mainly because I want there to be this limit. I still believe that challenging, thought-provoking processes should remain human. I also notice that team shrinking from phase to phase. According to the chart, the engineer is now the primary user and there really is no need for the other experts with modern systems. Good for the bottom line I guess … terrible if you want to have a beer after work on a warm spring Friday afternoon :-)
This was not intended to be a post on modeling. My goal was to highlight what I believe is a wonderful technical graphic created by one of the more insightful people I have ever met. And the graphic did its job … it got me waxing on fundamental issues of modeling automation. This is the essential quality of an exceptional graphic: it makes you think.
Incidentally, if you are interested in fresh perspectives on technical graphics and visualization, I highly recommend Edward R. Tuffte’s landmark book The Visual Display of Quantitative Information. One of the most enjoyable ways of spending an afternoon … lots of pictures!
For more info on Edward Tuffte’s book, The Visual Display of Quantitative Information