From CSULA CS Wiki
Jump to: navigation, search

Sarah, MiLK here. Your paper is good. It offers a good starting point for a definition of a complex-system. It collects actionable insights that can be used to engineer solutions to the problems that are associated with such systems. I would offer the following comments:

1. Your definition of a complex system has seven major elements. Elements 1 and 7, and elements 2 and 5 overlap substantially and might well be combined.

2. The distinction between system and environment is important - essentially your element 3 (the system is negatively entropic, self organizing).

3. Abbott uses the word "entailment" to capture the overlaps between conceptualizations at multiple scales of conceptualization that do not obey the laws of reductionism and determinism. Both he and you focus on "upward" emergence. I would note that the influence of what can be conceptualized at "higher" scales can have an analogous influence on "lower" scales of conceptualization.

4. cSE should not be viewed as displacing or replacing TSE. cSE should be used to address the realization of solutions to problems that are not amenable to TSE.

5. cSE involves intervention by engineers in the behavior and the form (both substance and structure) of a system that the engineers are a part of. I would substitute a notion like "span of influence" for the "span of control" familiar to system engineers.


--George McC 13:51, 10 January 2007 (PST)

Hi Sarah,

Unlike all the other papers that I am writing about, at least you have the advantage of knowing "where I am coming from" with my comments! That may or may not be a good thing.

I am most impressed, first of all, with the overall standard of what I have been reading as I have gone through the papers and with the wealth of good ideas. Congratulations on helping to bring it all together. I'm even more disappointed that I cannot join in the discussions face to face.

My major comment on your paper is that, perhaps, it is covering too much ground. As I was reading through it I kept thinking that there needed to be more time spent on each of the topics that you have produced a "whistle-stop" tour through. Certainly I could find much that I would discuss/clarify/even disagree with. I will limit myself to just a few points here.

  1. As I may have said to you before - I have a knee-jerk defensive reaction when I see the phrase system of systems - I do think that there are such things, but often it is used simply as a name for a "big system" and I do have to take issue with your issues that distinguish system of systems from a system on page 2. I would argue that all five, and certainly 2 and 4, are prevalent issues with many systems.
  2. I like the table comparing INCOSE, processes and NAS - most importantly it brings out that there are different classes of systems that we must embrace.
  3. Table 3 comparing TSE with CSE - I would see each of the rows representing two ends of a spectrum with a very fuzzy boundary along that spectrum that separates TSE from CSE - I wouldn't put them wholly at either end.
  4. I agree with MiLK's point 4 above - it is not desirable to throw away TSE.