I came across Conway's law on 9/18/20. The note has a personal significance and not necessarily for a would be visitor of this post. I will state it anyway. What's the noteworthy about Conway's law? Let the author, Melvin Conway speak for himself:
Conway’s law stated in 1968 says: Any organization that designs a system (defined broadly) will produce a design whose structure is a copy of the organization's communication structure.
I could not find any hint to this law (it's in my view a curious recurring fact and less of a law as laws are either ecumenical or subject to repeated empirical observations/testing.) in the PM BOK. Perhaps in future we could open up a small corner for Melvin this reference model of project management norms.
In his 1968 brief paper (here's a link) published in Datamation journal, Conway suggested a link, a map, or a bridge from the way people (in a design organization/department) communicate with each other to the way the product is actually put together. In a broad respect (in a wider view) this law says that if you see a design (a specification, a machine, etc.) it's an indication of the way of communication among its designers. The indication in Conway's words (and assessment) had a stronger (mathematical) requirement: that the communication across the design team was preserved in the structure of the configured design. We really need a mathematician (with enough insight into both design theory/practice and graph theory) to tell us if the technical term used by Conway does really hold, ie, the homomorphism. A homomorphism for visual depiction is like an arrow! It is a pointer from A to B which by so doing also induces a path and direction, 'from A to B'. I am not stating the graph theoretical notion of the homomorphism as it might scare away the non-mathematical reader.
Through practice and in my opinion, I am inclined to take sides with Conway and see enough of topical/historical/indicative evidence for his formulation. I see he had a fine and deep point about the bridging of two vastly (if not total separate) realms of sciences: the communication among people which is a social science topic, and the configuration of a design which subsumes a few disciplines under this name (system science, design theory, engineering design, and pattern theory). I love Conway's paper!
I began this post to suggest an extended version of his observation that there is a mirroring relationship between the design team collaboration and the performing design. I've been looking into this relationship for sometime. More on this later.
Now if there’s a direct relation between the
communication structure and the structure of the designed product of an organization,
could one look at this model the other way around and let's say design a communication structure that
could be copied in the designed product? It’s tempting to say yes, That's WWW! The communication on the Web
is based on the TCP/IP protocol. Jon Kleinberg depicts (Networks, Crowds,
Markets, 2010) the Web as a bow-tie structure with a strongly connect component.
As he says:
Taken as a whole, then, the bow-tie picture of the Web provides a high-level view of the Web’s structure, based on its reachability properties and how its strongly connected components fit together. (p. 391)
What about the (inter-nodes) communication structure, the TCP/IP. This is something that may not readily lend itself to such exercise. To be continued.
Ali
