Agile Sensemaking


I am giving a talk in a few weeks at it is about sense making and devops.

Sense making is not a new activity but perhaps a new buzz word. Sense making is how we understand the world in order. We can look at this in relation to say Fordist production. The assembly line worker does not have to understand the world. the assembly line worker does to have to understand anything beyond her own task. She does not even have to understand how this fits into the world. The decisions involved in optimizing a Fordist production line have to do with optimization and efficiency. It is an operations research problem. It does not have to make sense of the production line in relation to issues outside the production line, such as sourcing of rare earth elements, pollution caused by finished products, physical issues such as carpal tunnel caused by the products from the assembly line or even physical issues caused by working on the assembly line.

I am going to mention three types of decision making, top down, peer to peer and self-organized (self-soverign). Top down would be traditional hierarchical decision making. These sorts laws remind me of the deontological ethics or value ethics. There is a rule and then people have to follow the rule. Peer to peer decision making, or collective decision making, is sort of what we have in a democracy. However, in its nefarious moments it is like Bentham’s panopticon. These are the laws according to optimization… Mill’s greatest good for the greatest number. Self-soverign is that people make their own decisions that suit them. This is not a law, like Kant’s notion of the Kingdom of ends, where individuals are both subjects and rulers of themselves. For self-organization or self-sovereignty there is no ruling or subjectification – there is only acting with integrity. Every action is the right action. Self-sovereign could also be called agent based decision making. Since each agent is making a decision

How do we become self sovereign? Good Question.

But also what happens if we run processes that used be run according to ideas of hierarchy or peer-to-peer organization shift.

a) What are the benefits

b) What are the organizational shifts

The benefits would be a less fragile system. A hierarchy is fragile, a hierarchy also has a large infrastructure to support it so it is expensive, a hierarchy moves slowly it is not nimble. A peer-to-peer system is more resilient, but it too requires infrastructure, and the peer-to-peer nature does presuppose an architecture, a pipeline that has to be maintained. In an agent or self-organizing, there is no architecture, only the orientation of the agent.

How does this apply to something like a DevOps pipeline or DevOps operations. Well if we imagine the organization of a k8 cluster as either hierarchical or peer-to-peer, what would be an example of self-organizing. What would be an example of a software application that could sense make, as opposed to decision making?

I’ve been thinking about this today in related to this piece by Bonnitta Roy.