--George McC 06:58, 9 January 2007 (PST)
I have to admit that my first reaction when I read the phrase "system of systems" is to cringe - mainly because the majority of people who use the term seem to me to be using it to mean (simply) a complex system. Sometimes the use of the term goes beyond this, as in your paper, however I am still unconvinced of the efficacy of the term.
My own view is that what you are describing as a "system of systems" is a particular class of system - and I am unconvinced by any suggestion that a SoS is not a system. This is, of course, partly (completely?) a semantic problem. I view systems as transient things - for the most part changing size, shape, composition as and when necessary and being nested in a holarchy of other systems.
Be that as it may - I ask myself is there a purpose to why we need to identify SoS (as you define them) as distinct from systems. The reality is that many have now latched on to the phrase SoS - although I reiterate my initial statement that many misuse it - and therefore it is something that must be considered.
I can see that the common use (certainly within the defence domain) is that a SoS is a collection of procured items (systems) working together towards some common objective. The problem, as I see it, is where do you draw the line - is a communications system consisting of many switches, handsets, cables etc. a SoS? is a naval ship with its weapons systems, sensor systems, command system etc. a SoS? is the fleet a SoS?
If I were to be able to redefine the use of SoS (which I clearly cannot, since it has now become ingrained in the language) I would say that a SoS occurs where there are interacting/cooperating/competing systems that require to be considered in order to address a particular situation. This is recognised, although probably not intentionally, within initiatives like the UK MoD's lines of development (training, doctrine, equipment etc) - each of these is a competing/cooperating system and together they form a SoS.
I guess that (and this is the first time I have even articulated this to myself!) what I am saying is that the distinguishing feature of a SoS is that there may not be a common objective - rather there are several systems (each with their own objectives) that interact to produce an effect (which perhaps none of them wanted!).
Is this just a load of old twaddle or is there some sense in there somewhere?
-- Matt B: No it makes sense to me. SoS is about interactions between systems, and is often really just complex systems---I guess an example of a complex system which isn't a SoS would be a proton, and a SoS which I would call complicated, but not complex, is a car. But in general SoS is equivalent to a complex system. Your last point reminds me of the Tragedy of the commons or an over-full El Farol bar (sadly, my complex systems friends and I were unable to disturb the steady state as we were put into a separate beer garden away from the front bar). Your comments on the transience of systems are insightful.