A work in progress

For over a year now I have been playing with differ ways of representing the complex domain of Cynefin. As a result of that process I have been growing in my conviction that we framework for complexity and its three boundaries (to chaos, to complicated and to disorder) is key to working through intervention models beyond the generic safe-to-fail probe approach. I've got a whole moleskin notebook full of sketches on this and its still got a long way to go.

However over the last couple of days I started to play with one of the approaches around requirements capture and it seemed to work. Given that i am sharing it for comment with the very important qualification that this is, as the title says, a work in progress. That means it is unlikely to survive in its current form for long. This is also the two dimensional version, there is a three dimension one as well but I am still drawing that.

The dimensions here pick up in the 3Cs I briefly mentioned yesterday, but in the two dimensional version the main axes are:

• Coherence: the degree to which any need or requirement is structured/defined/understood
• Convergence: the degree to which different interest groups agree on the needs and nature of what is needed

Now in an ideal world we would want both convergence and coherence of requirements. That is to say we want people to agree on what is needed, and we what the subject of that agreement to be firm enough that there is little ambiguity. That will also mean that there is little fragmentation of said requirement and we can just get on and do it. That is the dark green section on the framework and represents an ideal state, namely something we will never achieve in reality on other than rare occasions IF (and this is critical), IF we attempt to start there. I'll come back to that in a minute, but lets first of all look at inauthentic spaces or danger areas:

• Pattern Entrainment – High convergence, low coherence
  There is lots of agreement about what is needed but little structure to the definition, possibly because there is clear agreement. This is where assumptions can be followed without proper testing or consideration of the need. Its often the area where past practice is retaining its attractiveness long after its applicability has fallen to changing context.
• Power Play – Low convergence, high coherence
  A very structured requirement with little agreement between factions. The most powerful faction will win with all the possible downsides in lack of commitment, failure to scan enough a diverse set of needs and so on. A variation on this is the maverick who does know but is isolated by fractious politics.
• Strong belief – Low/Medium convergence, low coherence
  I'm not sure of the name here, but this is the domain of different factions with similar power resulted in a fractured and disjoined position. This is one of the issues that techniques such as SNS are designed to resolve; by allowing different groups to work in parallel with interaction, conflict can be resolved through action not dialogue.
• Confused – Low convergence, low/medium coherence
  We've got some structure in the need but we don't even have factions fighting between the options. Individuals have needs but there is no clumping or links between those individuals. It's a mess with few patterns or structures that we can do anything with.

Now just to make this very clear – I have produced some examples here but there are probably better ones. I am thinking aloud here and reserve the right to delete/change the above without more than a token apology for their suitability.   

Now the green represents a valid range or variance and the implication is that requirements need to move to the coalescence line (a sense of Ashby's Law here) and then move up or down the diagonal. As they move bottom right constraints start to disappear and the position becomes increasingly chaotic. As it moves top right the constrain increases as it shifts from complex to ordered.

The idea (moving forward) is to look at a diagnostic (where are we) which would then provide a typology of interventions to bring things back on line, and ultimately on track.

As I said its a work in progress and an early stage at that. Comments or questions welcome here or by email