I was watching 007: Skyfall last week. I think it just came out. Spy movies are one of the main homestays of the word “protocol”. When something goes wrong – initiate the protocol. When the protagonist goes rogue, it’s a breach of protocol. This sort of film tends to give preferential treatment to those who break the rules of the establishment in order to save the day, stop the bad guy, and get the damsel in distress.
The MI6 bureaucrats and their lackeys are cast as stiff, nerdy, and fragile by-the-book operators. Elite spies on the other hand are the epitome of free will and valor. As Merve Emre explains in Bureaucratic Heroism, it’s not surprising that this portrayal resonates with me and other people my age. We grew up watching James Bond, Mission Impossible, Die Hard, Iron Man, etc. In order to win, the hero had to fight both the enemy and a bloated administration.
There is an alternative sub-genre of action movies that Emre also points out. One where it is a high-functioning bureaucracy that gets the job done. In this world, the saviors are not so much men and women with unnaturally free wills. Rather, the protagonist is the rules of the institution – protocols – carried out by dogged agents of the state.
I think this version is a pretty realistic version of how things work in the real world. “The wheels of justice turn slowly, but grind exceedingly fine” has a lot of truth to it. For the past several hundred years, the organs of democracy have succeeded in tackling many big problems. These include establishing a rule of law, consumer protection rights, public healthcare systems, tax enforcement, and environmental protection.
But I don’t find the bureaucracy to be a good hero. Call me a romantic, but I like to imagine that it is individuals that change the world. So, if we work backward to find a character to attribute this bureaucratic heroism – who will it be?
The protocol maker. The hidden, unnamed actor in every spy movie. The inventor of these emergency-, code red-, secret-, ghost-protocols is never mentioned. But the protocols exist, and they were written by someone. It’s probably more accurate to say that these protocol makers are gardeners. Protocols, as patterns of human behavior, develop more like plants than construction projects.
There are two types of protocol makers: planters and gardeners. Planters, who write the initial protocol, have to spread it to a sufficiently big enough plot (a human network). They must onboard people, teach them the protocol, and give them the tools to pass it on. Then the planter quietly fades into the background. Gardeners tinker with existing protocols in order to improve them or resurrect them from a state of hibernation. In the spy movie context, a gardener might get named, but there’s a 50/50 chance that they’re cast as an uptight manager.
Whether planting or gardening, these actors face a trilemma. Project managers face an inherent trade-off between scope, quality, and schedule. Scrum masters compromise between value, budget, and deadline. In data science, the CAP theorem states that a data store can only provide ⅔ of the following: consistency, availability, and partition tolerance.
Similarly, a protocol can’t perfectly meet all of the following criteria:
Comprehensive: The protocol can be activated in a wide variety of contexts.
Learnable: Onboarding new users to the protocol is fast and low-cost.
Secure: It’s difficult to hack the protocol by faking compliance, overloading, editing, etc.
The more comprehensive a protocol is, the easier it is to breach and harder it is to learn. The protocols for enforcing the American tax code, for example, are extremely comprehensive. But for that reason, they can be easy to hack and are obviously insanely difficult to learn.
Learnability comes at the cost of context-flexibility and security. Simplistic protocols break upon contact with unusual contexts and edge cases. At worst, they can be overtly harmful when applied in the wrong contexts. A medical protocol promoting the consumption of electrolytes (incl. sodium) could hurt someone with high blood pressure. An easy-to-learn protocol is also easy to mimic, and its internal logic is legible enough to be rewritten.
Protocols stay secure either by i) increasing the barriers to entry (learnability) or ii) limiting the range of their application in order to reduce side effects. Clearance levels are a way of increasing security that reduces access, and therefore learnability. Geographic jurisdictions can be used to constrain the range of applications.
In their initial stages of life, projects, teams, data stores, and protocols are malleable along their trade-off axes. But as they age they become more rigid. It’s like distributing skill points on a videogame avatar. If you’re only level 2, it’s easy to restart and redistribute. But if you’re deep into the game, you’re basically locked into a given strategy because you’ve irreversibly invested in certain traits.
