On the path to improving the reliability of a given system, the oscillation between two modes of work is often necessary. The first mode here is that which is directly affecting the change on a system, and the second describes that which is improving the method by which change is made on a given system.

These are references for a mental model that could be applied to a system as simple as a lever, or as complicated as an automobile manufacturing plant. The systems I tend to work most closely with are computer and network systems, but this is likely applicable for other sorts of system operations where humans are concerned.

Mode 1

During the period of time when a human operator must be interacting directly with a system, there are some observations worth making.

First, there is no multiplier or leveraged gain for each unit of effort. This could be viewed as the least valuable work, due to its lack of scalability, but only in contrast to the second mode. No artifact is produced on which the next artifact is built, or leveraged against. For all intent, this is a linearity.

For a reference, a manual transmission in a car is a linear amount of effort to work accomplished. The human operator assesses the state of the system to determine that a gear change should be made, and then takes the appropriate actions to move the transmission from one gear to another. Nothing beyond that one gear shift is achieved, and thus the next occurrence of a gear change will require yet more human effort equal to that which preceded it.

A second observation is that the direct nature of operating the system places a certain back-pressure or feedback from the external inputs to the system. This pressure is transfered to the human operator through observation, feel, and stress. The operator must be free enough to react to the ever changing inputs to the system, so as to make the necessary adjustments to keep the system running efficiently. Free here in the sense to avoid competition with other objectives for the time and attention of the operator. If a part breaks, or a spontaneous combustion causes a fire, or there is an object in the road ahead of a vehicle, the operator must be able to either divert attention to resolve the obstacle directly, or signal for help.

A third consideration is of the experience that the operator gains while operating in this mode. Due to the felt feedback from the system, the operator is often acutely aware of potential risks to the operation or the challenges with keeping the system running optimally, which increases stress on the person.

In the conditions above, it is the operator that is the constrained resource due to the direct nature of the work. Deadlines in this mode are of auxiliary consideration, since, in order to keep the system running the operator is only able to move the system at a pace congruent with the resistance of external inputs and within the limits of the system.

Mode 2

The second mode is the effort directed at scaling the system, or improving the efficiency of running the system. The product or output of this work is leveraged against prior work. Perhaps too, this work will pave the way for yet future work. Considering this trajectory of past, present and future can provide insights into strategic decisions that will need to be made about the operation.

This is the work of codifying the operational intelligence into a mechanical operator. The humans review a known set of conditions for which the mechanical operator must be equipped to react as the running system exhibits feedback from the externalities. As new failure conditions of the system are are known, those conditions are met with further development to improve the mechanical operator’s ability to handle the conditions as they arise.

To use the above example of an automobile, an automatic transmission would be the corollary here. It is the mechanical operator that assesses the system state and responds to changes in inputs to determine the appropriate gear and posses the correct mechanical apparatus to make the adjustments, shifting gears.

The engineers who developed and designed this system were working in the second mode to achieve the mechanical operations that shipped with the final product.

During development of the mechanical operator, the output of work is perhaps slower to prove value, though its non-linearity out-competes the output from the previous mode given sufficient development duration and proper decision making about which developments to prioritize. As development continues, this new mechanical operator will take more tasks from the human operator over time with improvements over time, though each task requires the human designers to implement.

Project management

In the case of businesses, projects for this body work are often thought about in advance among a group of individuals, with a good deal of planning and design ahead of time. The expected results are often targeting some aspect of meeting business objectives. Often the constraint of deadlines here due to business cadence means that there is some amount of rush that needs to take place, and sacrifices must be made to either the quality of work delivered, or the scope of the work delivered.

Finding balance

It is when these two modes of work compete for attention with one another that this becomes a problem for an individual for time or attention. The two objectives are in opposition to one another due to the limited human resource. One cannot both reach a deadline while operating a system who’s inputs to from external sources are unknown or not under control. Operating takes as long as it takes or risks to the functioning of the system are at stake. These risks vary depending on the system of operation, but could be severe enough to impact the well-being and physical safety of people involved.

Working in teams

These modes of operation should be divided among humans and rotated periodically to avoid burnout from the endured stress of receiving feedback from the system.

Given the linearity of the first mode, limiting the number of people operating in this mode seems prudent. Though in an on-call situation, some amount of human resource must be dedicated to responding to issues as they are detected in the running system.

When working with a large system, the domain of operation may be further divided, so that the oscillation happens between the first mode and a more narrow area of focus in the second mode. This is to allow the operator to become familiar with a smaller set, and proficiency in that area is gained much faster. Rotation here could happen here too, though perhaps at a slower pace.

Feedback

As the effort from operating in the second mode feeds into the running system to improve the capabilities of effort over time, the operator working in the first mode is able to achieve more output given the level of effort. It is in this way that the second mode is leveraged to support the first mode. The amount of time spent working in the first mode may not decrease over time, but the capabilities and outputs may increase given the aide of a mechanical co-operator.

Ode to the operators

None of the above is to suggest that one should only spend time working in the second mode due to its perceived higher value. The primary reason for a system is to be operable at all times. The aim here is to improve the operation over time with respect to efficiency, time, and attention, allowing the operator more control, flexibility, and capability when operating in the first mode becomes necessary. The perspective over time merely shifts to one with more capability, and one of perhaps more altitude, while the details are perhaps coded deeper within the stack and are assessed outside the pressures and time constraints of operating the system directly in the fist mode.

Much more in this area is worth exploring, but perhaps another time.