Thursday, January 25, 2007

Systems Engineering Start

This is a blog for my notes on the MetaOlin independent study that six of us are doing in Fall 2006. We will be looking at Olin from a bunch of different perspectives, our first module being Systems Engineering (for the "10k feet up" view, as Brian Bingham said). Brian gave us some readings to kick things off.

Robert W. Lucky, Bozos on the Bus (IEEE Spectrum, 1996)

One sentence summary: We're blindly wandering in the dark and have no idea of what's coming ahead... but neither does anyone else, and this levels the playing field.

Question: Is the advantage in the new world system (whatever the heck that means) and the creation of new systems (products, etc.) skewed towards the young? Since young people tend to be much more used to "being bozos" on account of not knowing enough to be much of anything else, we're apt to adapt better to a world where everyone's thrown into bozo-hood, much like being blind during a nighttime power outage. I have a hard time believing this; experience and the wisdom of years is usually transferrable to different situations.

Daniel Hastings, The Future of Engineering Systems: Development of Engineering Leaders

Summary: Systems are all around us - education, healthcare, government, etc. These systems are becoming increasingly technologically enabled (telecommunications, the internet...), but most engineers haven't learned how to work with systems of this magnitude that also interact with nontechnical, more sociological things. Since there are many possible views of the same complex system, Systems engineers have to be able to synthesize many different perspectives at once.

Systems engineering typically focuses on things with the following properties

  • Technologically enabled - there are components that require solid engineering backgrounds.
  • Large scale & complex - wide-reaching, far-ranging, and plenty of parts.
  • Dynamic uncertainty - things should be changing in this system, and it should not be strictly predictable (fairly easy to get this when the system is large scale and complex).
  • Interaction with nontechnical factors - there are people in the world, and we need to account for them.
  • Emergent behavior - the whole is greater than the sum of its parts; some things will come out that you never predicted.
I was particularly fascinated by the paper's insistence that we need more rigor and mathematics and quantitative modeling!!! in systems education. It's almost as if they're trying to make sure that systems engineering is a "real" discipline, not a "fuzzy, soft thing" that "real engineers" will dismiss as hand-wavy.

This reminds me of a conversation I had with MIT anthropology professor Dr. Susan Silbey about "engineer's arrogance," which is what UOCD at Olin tries to cure. We assume that since we know about technology, and we're people, that we can make technology for people without learning about them. We don't need humanities! We just know what to make.

Anyhow, as much as I love math, I hope that systems doesn't get sucked into the "if it isn't quantitative, it isn't real engineering" trap. There are many ways to solve a problem, and numbers are only one.

The other part that I was struck by was the repeated echoes of "this discipline is not mature." That's exciting. As Lucky's paper put it, we're all bozos on this bus. Bill Strachan from IBM told me that systems engineering right now is around the same place that computer science was a generation ago - it's drawing on all these other disciplines, but it's more than the sum of their parts... in other words, systems engineering is in itself a system. It fits all the above criteria, after all.

No comments: