REPLY Society Strategy Development (SD6971)

[SDMAIL Rick Kossik]

Posted by  "Rick Kossik" <RKossik at goldsim.com>

As an engineer who has built dynamic models of complex systems for over 
20 years (using, in some cases, SD concepts and constructs), I find much 
of the discussion on this board regarding finding ways to increase the 
use of SD in the real world a bit confusing.  Although I recently 
joined the SDS (and attended and thoroughly enjoyed the Boston 
conference), I am really an outsider:  Although I am from MIT (Civil 
Engineering, not Sloan), I learned dynamic modeling outside of the SD 
community (through spending 20 years as a consulting engineer), and have 
never taken a formal SD course.  As such, I hope that some comments from 
an "outsider" will add some value to this discussion.  Many of the ideas 
below have been mentioned or at least alluded to by others, but I 
thought it would be worthwhile to reiterate them here, as an "outside 
voice".

It seems to me that many of the suggestions for advancing SD tend to 
revolve around training and graduating more SD modelers from system 
dynamics programs (e.g., this was mentioned a few times at the 
conference too).  To an outsider like me, this seems analogous to 
suggesting that we can solve more engineering and economic problems in 
the world by graduating more mathematics professors.  Mathematics 
professors do not generally solve applied engineering and economic 
problems (as they typically don’t have the domain-specific expertise to 
do so).  Rather, engineers and economists who understand the use of 
calculus and other mathematical tools solve those problems.

Putting the focus on training more SD experts seems to imply that 
solving complex engineering, business and social problems requires 
nothing more than expertise in system dynamics, and a couple weeks 
discussing the problem with discipline experts.  Again and again in this 
forum I read about how difficult it is to master SD.  However, the 
(probably unintentional) implication is that SD takes years to master; 
but disciplines like economics and environmental engineering (to which 
you want to apply SD) can be mastered quite quickly by a seasoned SD 
modeler by simply talking to some experts a few times.  Not only is this 
unrealistic, I think it can also come across as arrogant and as such, 
can have the affect of producing push back from discipline experts, 
resulting in a decreased use of SD.

SD is a powerful and useful tool that is applicable within many 
disciplines.  However, to effectively use any modeling tool in a 
particular discipline requires a thorough, fundamental and sometimes 
subtle understanding of the system and processes being modeled.  I do 
not believe that this can be acquired in a couple meetings with domain 
experts. That is, what we need is not more system dynamic modelers who 
know something about economics, but more economists who understand 
system dynamics.  Therefore, I believe that if you want to increase the 
use of SD, instead of focusing on training and graduating more SD 
modelers, you should put all of your focus on encouraging various 
disciplines to incorporate a semester or two of SD training into their 
curriculum. 

As an aside, it is worth reminding everyone (as has occasionally been 
done by others here) that SD is just one of many modeling and simulation 
tools that any particular discipline requires.  Many real world dynamic 
systems simply cannot be modeled using SD (e.g., many systems cannot be 
realistically described using only ordinary differential equations).  I 
mention this because someone trained primarily as a SD modeler may tend 
to see every problem as a SD problem (if all I have is a hammer, 
everything looks like a nail).  However, a good engineer, for example, 
carries a large number of tools in his toolkit, and applies whichever 
one is most appropriate to the problem at hand.

You will be successful if in 2028, the majority of the conference 
attendees do not think of themselves primarily as system dynamics 
modelers, but as economists, engineers, scientists and other analysts 
who sometimes (but not always) use system dynamics in their discipline, 
and have interesting and successful SD applications to describe.

Regards,

Rick Kossik
Posted by  "Rick Kossik" <RKossik at goldsim.com>
posting date  Thu, 24 Apr 2008 12:15:29 -0700