As an engineering post-graduate student a few years back, and now an engineer working in industry, I have often had this realization very strongly - that most problems we "adults" work on are about "scale" and managing the "chaos" in scale.

By "adults", I mean when we move beyond our graduation and towards post-graduation or industry, where we finally start to solve some real-world problems.

I will take a few examples both from industry and academia:

  • While doing research in robotics (multi-body dynamics) as a graduate student, I realized that, however challenging the field may have looked to me before, what I really was doing was solving some algebraic equations. As a child, I used to use simple algebra when problems had smaller number of equations; now I was using Linear Algebra to deal with the large number of equations. The problem really was no more challenging than before, only the scale had changed, and hence I had to learn the tools to deal with the scale.

  • Some of the most important subjects that are recommended to all applied engineering graduates are Linear Algebra and Statistics. Both these subjects really train you to how to deal with "scale".

  • The concept of "reproducibile research" is really about "managing the chaos" in scale. As more and more people are doing research, it is becoming imperative that academicians document their work meticulously and make it reproducible at any later stage by anyone. It's no more about simply solving problems and moving on.

  • Currently, working in a startup designing an IoT device, most of the design decisions we take are about the "scalability". We deployed a few devices in some pilot projects and it was easy to handle 10-20 devices. But as the number started to grow, it became so difficult to maintain the devices, that most of the innovation started to be driven by this factor.

I feel that the puzzling nature of problems (as it used to be in childhood, e.g. the thrill of solving Mathematics olympiad problems) gets diminished when you are into solving real engineering problems. The real puzzle (and the thrill!) lies in "scale".

Can someone shed some light on this? Has someone (some engineer/scientist/philosopher) written something on this topic/issue? Moreover, how true do you feel this realization really is? It may be a transient realization because of the current nature of work that I am doing in the industry.

Further, I'm not sure if this forum is the right place to ask this question. Kindly suggest me a different forum, if the question is too off-topic.

  • 1
    The importance and impact of how things scale is a major topic in network science (especially with power law distributions, "scale-free" networks), and this is actually a very big topic in tech generally - it's why the words "at scale" gets thrown around a lot by tech companies, and where "big data" comes from. At some level of scale you get new problems you never encounter before (like what you do when the data you want won't even fit on disk, much less in memory, as with streams). I think Herbert Simon is one of the earliest people I encountered who talked about these things at length. – BrianH Feb 20 '19 at 21:22
  • 1
    This will probably be considered too discussion-forum-esque for here, but you might be interested in Simon's book for a foundational reference on this sort of philosophical/mathematical/scientific thinking: mitpress.mit.edu/books/sciences-artificial – BrianH Feb 20 '19 at 21:24
  • Thanks @BrianH for suggesting Simon. I will read about him. And his book looks interesting! Could you suggest some forum where I could post this, if it's too forum-esque for StackExchange. – shivams Feb 20 '19 at 21:29
  • 1
    No, I don't think this it is generally the case that most real-world problems are fundamentally about "scale". For example, it is also common to encounter problems that are more accurately described as being "estimation" problems or "optimization" problems, which are separate issues from "scale". You use the term "chaos" in a colloquial sense, but there are real-world problems that deal with "chaos" in a mathematical sense that appears in dynamic models of systems and is not fundamentally an issue of "scale" (small systems of few equations can still exhibit chaotic behavior). – Matt Feb 20 '19 at 21:37
  • 1
    Most engineering problems center around schedule, cost, and performance. "Scale" may enter in to the tradeoffs between those 3, but is not a fundamental issue. – Jon Custer Feb 21 '19 at 0:06
  1. I would be careful about the use of the term "scale". A much more natural interpretation of that word in the context of engineering would be physical scale (e.g. how a model in a wind tunnel correlates to an airplane). Or issues with heavy or large (or small) objects. I got a little confused on what you emphasize, but think it is basically computational intentisity (e.g. finite element analysis).

  2. While I think you identify a reasonably common theme, I would not be so strong as you are in saying this is what engineering jobs are all about. For example there are huge amounts of people doing engineering in basic HVAC or the like that don't do intense FEA type stuff. I would just be careful about letting your particular experience prejudice you to that being what all engineers do. I know a LOT of them that don't do ANY linear algebra on the job. Instead they add up loads and then trace their finger along a nomograph in the Ingersol Rand catalog to pick a pump. (Not to mention other aspects of the job, layout, drafting, inspection in the field, cost estimation, etc.) In addition, I think you will find if you look at different fields of engineering (outside of the stereotypical mechE) that the tools and tasks differ. E.g. pharma process engineering, naval architecture, electrical engineering (power versus micro), chemE (refineries), petroleum engineering (completions).

| improve this answer | |

Not the answer you're looking for? Browse other questions tagged or ask your own question.