Note the similarity between Hayek's passage above and the quotation below that Klein (and co-author) selects from Schelling to motivate the discussion of mutual coordination.
Yet in their 2009 paper, Klein and Orsborn introduce the above Schelling quote to distinguish the new usage of the term coordination from the older meaning that economists such as Hayek (and others at the London School of Economics) had had in mind.
To repeat my earlier claim, I suggest that if a randomly selected Austrian (who was unfamiliar with Klein's treatment) were asked whether Hayek's knowledge papers had to do with coordination in the sense of an interior designer planning the color scheme of a living room versus the sense of friends synchronizing their plans to meet up for a movie, then it is very likely that the Austrian would say Hayek's usage lined up with the second sense (i.
Rather than claim (as Klein does) that Hayek focused on concatenate coordination (in which individual actions must fit together in such a way to yield a pleasing outcome to a superindividual observer), whereas Schelling focused on mutual coordination (in which an individual must adjust his action in light of what others are expected to do), instead I would say the critical distinction between the two camps is this: Hayek focused on the tremendous difficulties in achieving equilibrium at all, whereas Schelling (and many modern neoclassicals) assume away this real-world problem and instead focus on choosing from among possible equilibria in stylized games that are much simpler than the actual economy.
The term coordination has had an extensive and varied history in economics, as Klein and Orsborn (2009) document.
Unfortunately, after spending so much time outlining his suggested taxonomy (originally between coordination and metacoordination but now between mutual coordination and concatenate coordination), Klein fails to distinguish the essential differences between the Hayekian and Schelling uses of the term.
The coordination layer intercepts messages among actors and applies coordination constraints on the messages.
From the perspective of a coordinator, a role enables the coordination of a set of actors that share the same static description of behaviors without requiring the coordinator to be aware of the individual actors in the set.
To comply with the separation of concern principle, we categorize these actors into two types: computation actors that capture system computation concerns, and coordination actors that abstract system coordination concerns.
Events are special messages that are atomically dispatched on coordination actors.
This requirement guarantees that coordination constraints are applied on related messages before these messages are dispatched on computation actors.
To maintain coordination transparency and avoid interfering with the computation actors' functionalities, coordination actors are not allowed to generate or send messages to computation actors.
