Where the frontier of science once was is now the centre.
― Georg Christoph Lichtenberg
I’ll just say up front that my contention is that there is precious little excellence to be found today in many fields. Modeling & simulation is no different. I will also contend that excellence is relatively easy to obtain, or at the very least a key change in mindset will move us in that direction. This change in mindset is relatively small, but essential. It deals with the general setting of satisfaction with the current state and whether restlessness exists that ends up allowing progress to be sought. Too often there seems to be an innate satisfaction with too much of the “ecosystem” for modeling & simulation, and not enough agitation for progress. We should continually seek the opportunity and need for progress in the full spectrum of work. Our obsession with planning, and micromanagement of research ends up choking the success from everything it touches by short-circuiting the entire natural process of progress, discovery and serendipity.
In my view the desire for continual progress is the essence of excellence. When I see the broad field of modeling & simulation the need for progress seems pervasive and deep. When I hear our leaders talk such needs are muted and progress seems to only depend on a few simple areas of focus. Such a focus is always warranted if there is an opportunity to be taken advantage of. Instead we seem to be in an age where the technological opportunity being sought is arrayed against progress, computer hardware. In the process of trying to force progress where it is less available the true engines of progress are being shut down. This represents mismanagement of epic proportions and needs to be met with calls for sanity and intelligence in our future.
If you find that you’re spending almost all your time on theory, start turning some attention to practical things; it will improve your theories. If you find that you’re spending almost all your time on practice, start turning some attention to theoretical things; it will improve your practice.
So how do we get better at modeling & simulation? The first thing is mindset; do we think going into our work, my goal is to make this thing good? Or state of the art? Or simply how can I make it better? The final question is the “right” one, you can always make it better, and in the process the other two questions will get answered. Too often today we never get into the fundamental mode of simply working toward continual improvement as our default mode of operation. The manner of energizing our work to do this is frightenly simple to pursue, but rarely in evidence today.
The way toward excellence, innovation and improvement is to figure out how to break what you have. Always push your code to its breaking point; always know what reasonable (or even unreasonable) problems you can’t successfully solve. Lack of success can be defined in multiple ways including complete failure of a code, lack of convergence, lack of quality, or lack of accuracy. Generally people test their code where it works and if they are good code developers they continue to test the code all the time to make sure it still works. If you want to get better you push at the places where the code doesn’t work, or doesn’t work well. You make the problems where it didn’t work part of the ones that do work. This is the simple and straightforward way to progress, and it is stunning how few efforts follow this simple, and obvious path. It is the golden path that we deny ourselves of today.
The reasons for not engaging in this golden path are simple and completely and utterly pathological. The golden path is not easy to manage. This golden path is epitomized by out of the box thinking. Today we prize in the box thinking because it is suitable for management, and strict accountability. This strict accountability is the consequence of societal structures that lack trust and implicitly fear independent thought. Out of the box is unstructured and innovative eschewing management control. As such we introduce systems that push everything inside the proverbial box. Establishing results that are predictable has become tantamount to being trustworthy. Out of the box thinking is dangerous and the subject of fear because it cannot be predicted. This is the core of our current lack of innovation, and the malaise in modeling & simulation.
The element of thinking that is missing from how things currently progress is a sense of satisfaction about too much of what has driven the success of modeling & simulation to date. We are too satisfied that the state of the art is fine and good enough. We lack a general sense that improvements, and progress are always possible. Instead of a continual striving to improve, the approach of focused and planned breakthroughs has beset the field. We have a distinct management approach that provides distinctly oriented improvements while ignoring important swaths of the technical basis for modeling & simulation excellence. The result of this ignorance is an increasingly stagnant status quo that embraces “good enough” implicitly through a lack of support for “better”.
There seems to be a belief that the current brand of goal-oriented micromanagement is good for technical achievement. Nothing could be further from the truth; the current goal based management philosophy is completely counter-productive and antithetical to good science and achievement. It leads to systematic goal reduction and lack of risk-taking on the part of organizations. A big part of this is the impact of the management style on the intrinsic motivations of the scientists. Scientists tend to be quite easily and intrinsically motivated by curiosity and achievement while the management system is focused on extrinsic motivation.
The test of a man isn’t what you think he’ll do. It’s what he actually does.
― Frank Herbert
We end up undermining all of the natural and simple aspects that lead to productive, innovative excellence in work, replacing these factors with a system that undermines what comes naturally. All of this has a single root, lack of trust and faith in the people doing the work. Without rebuilding the fundamental trust in providing intrinsically motivated and talented people a productive environment, I fear nothing can be done to improve our outcomes and grasp the excellence that is there for the taking. People would gravitate toward excellence naturally if the management would simply trust them and work to resonate with people’s natural inclinations.
Modeling & simulation arose to utility in support of real things. It owes much of its prominence to the support of national defense during the cold war. Everything from fighter planes to nuclear weapons to bullets and bombs utilized modeling & simulation to strive toward the best possible weapon. Similarly modeling & simulation moved into the world of manufacturing aiding in the design and analysis of cars, planes and consumer products across the spectrum of the economy. The problem is that we have lost sight of the necessity of these real world products as the engine of improvement in modeling & simulation. Instead we have allowed computer hardware to become an end unto itself rather than simply a tool. Even in computing, hardware has little centrality to the field. In computing today, the “app” is king and the keys to the market hardware is simply a necessary detail.
To address the proverbial “elephant in the room” the national exascale program is neither a good goal, nor bold in any way. It is the actual antithesis of what we need for excellence. The entire program will only power the continued decline in achievement in the field. It is a big project that is being managed the same way bridges are built. Nothing of any excellence will come of it. It is not inspirational or aspirational either. It is stale. It is following the same path that we have been on for the past 20 years, improvement in modeling & simulation by hardware. We have tremendous places we might harness modeling & simulation to help produce and even enable great outcomes. None of these greater societal goods is in the frame with exascale. It is a program lacking a soul.
The question always comes to what am I suggesting be done instead? We need to couch our overall efforts in modeling & simulation in supporting real world objectives. Something like additive manufacturing comes to mind as a modern example that would serve us far better than faster computers. We need to define a default attitude that progress is always possible and always something to be sought. Unfortunately, this more sensible and productive approach is politically untenable today. The real problem isn’t intellectual or bound in thoughtful dialog, but rather bound to a deep lack of faith and trust in science and scientists. We have poorly thought through programs focused on marketing and micromanagement as a direct result. Progress be damned.
There is a very real danger present when we suppress our feelings to act on inspiration in exchange for the “safety” of the status quo.
We risk sacrificing the opportunity to live a more fulfilling and purpose driven life. We risk sacrificing the opportunity to make a difference in the lives of others. We risk sacrificing the beautiful blessing of finding a greater sense of meaning in our own lives.
In short, we run the very real risk living a life of regret.
― Richie Norton