A calculation that divides a waiting list by the seats of a classroom answers a question about space. Whether the same calculation answers a question about time depends on things the division does not contain: how long each group takes, whether groups can run side by side, and who goes first. This paper studies one set of decision records in which a small calculation of exactly this kind was executed, kept and replayed, and in which the records then speak of schedules, deadlines and pace. The question is what such a record takes its calculation to have settled, and what it leaves open.
1The question and the records
The records come from AEGIS, the safety and protection studio of the RISCON family, which publishes several thousand authored decision cases. Every one of them is a synthetic setting: AEGIS marks each case realistic synthetic and calls each record an authored question seed whose judgment mark is the author's record, not a factual basis from public records. Within them AEGIS keeps a calculation programme of cases that each execute one small arithmetic rule on figures the author set, and keep the input, the output, a hash of the input and a replay that ran the rule again. The eleven cases read here are every case of that programme whose rule counts training cohorts against seats. Each one sets a number of people who must be retrained because their posts are changing, the seats of one training cohort, a period for the change and a period for the training, and asks how the people should be divided. They range from a posting change of non-commissioned officers and maintenance staff retrained on new equipment to refresher training for reserve officers and a move to the cyber defence field. The posts, seats and months are the author's assumptions. Nothing in this paper is a finding about any real organisation, and nothing in it says how any training should be planned.
The research question is asked of the record set, not of the world. When a decision record turns a number of people waiting for retraining into cohorts by an executed calculation, what does the calculation settle, what does the record leave to time and to the order in which people go, and where does the record's own wording part from its own figures?
2The contribution
The paper makes four claims about these eleven records. First, the executed calculation settles one thing exactly: the people due, divided by the seats of a cohort and rounded up, give the number of cohorts. Every one of the eleven recommendations states that number, in figures or in words, and so rests on the run. Second, the calculation's only measure of time compares the end of the first cohort with the end of the transition period. It gives no time for the last cohort. Two records, whose own figures make the cohorts fit their deadline exactly when run one after another, say that the whole schedule fits; the other nine speak only of the number of cohorts, while their recommendations speak of no gap, of balance or of finishing the transition. Third, who goes first is the thing most often named as unknown, in seven records, and no recommendation sets an order; an order appears only among the alternatives. Fourth, in a few places the words of a record and its figures part: a situation that calls six times the seats double the seats, a condition for reopening that the record's own figures already meet, and options for more seats that, at the record's own figures, leave the number of cohorts as it was.
The contribution to the literature is a reading of where one designed record set puts the edge of its calculation. The debates gathered below argue about whether policy is best seen from the decision down or from the point of delivery up, about the discretion and rationing of those who deliver a service, and about how a number waiting, a rate and a time are tied together. This record set gives a small, fully visible instance in which a decision is framed from the top by an exact count of seats, while the facts the bottom would supply, the order of the queue and the time it takes to clear, are left to inputs the records say they do not have.
3The debate
The first strand concerns how the carrying out of a decision should be studied. Paul A. Sabatier, Top-Down and Bottom-Up Approaches to Implementation Research: A Critical Analysis and Suggested Synthesis (1986), reviews the implementation literature of the preceding fifteen years with emphasis on the relative strengths and weaknesses of the top-down and bottom-up approaches. It argues that the time-frame used in most implementation research misses many critical features of public policy-making, and proposes a framework for examining policy change over a much longer period that combines the best features of the two. This paper takes from it two things: the contrast between a view from the decision and a view from where the decision is carried out, and the warning that the time-frame used in most implementation research misses many critical features of policy-making, which this study reads more generally as a reminder that the period an analysis covers limits what it can see. The records here look over a short period by construction, as section 1.2 shows.
Benny Hjern and David O. Porter, Implementation Structures: A New Unit of Administrative Analysis (1981), argue for a unit of analysis wider than a single organisation. Programmes, on their account, are implemented by clusters of parts of public and private organisations, and the implementation structure is the administrative entity that those who carry out a programme use to accomplish its objectives; an analysis rooted in one organisation is inadequate to make sense of the activities around implementation. This paper reads the records' needed inputs through that claim: what the records say they lack comes from several places at once, a roster of who wishes to move, the seats of a training site, a schedule of equipment, and the instructors.
Richard E. Matland, Synthesizing the Implementation Literature: The Ambiguity-Conflict Model of Policy Implementation (1995), finds the field split into top-down and bottom-up schools and reconciles them by the ambiguity and the conflict of a policy. Where both are low, the model calls implementation administrative; where ambiguity is high and conflict low, it calls implementation experimental. This study reads the eleven records as written in the first of these registers: the goal is not in dispute, the means are a count, and the count is exact. The reading is this study's own, and the point of it is the remainder. What the count does not fix, the time of the last cohort and the order of the queue, is where the records' own uncertainty sits.
Laurence J. O'Toole Jr., Research on Policy Implementation: Assessment and Prospects (2000), assesses a field that no longer framed the core question of public management, finds modest to moderate progress on a number of fronts, and argues that work on questions close to implementation continues under other labels and in unusual places. A set of authored decision records that schedules retraining is such a place: it carries a decision down to the seats of a training site without using the word.
The second strand concerns those who deliver a service and the queues they manage. Peter Hupe and Michael Hill, Street-Level Bureaucracy and Public Accountability (2007), start from the concept of street-level bureaucracy coined by Michael Lipsky and from the relative autonomy it gave to professionals at the front line. They add that such professionals work in a micro-network of relations in varying contexts, and that the conception of governance adds the multi-dimensional character of a policy system as a nested sequence of decisions; on this ground they rethink how those at the street level are held accountable. Lars L. G. Tummers, Victor Bekkers, Evelien Vink and Michael Musheno, Coping During Public Service Delivery: A Conceptualization and Systematic Review of the Literature (2015), describe how frontline workers adapt to stress, for instance from high workloads, by coping, using practices such as breaking or bending rules for clients or rationing services, and show from a systematic review that the type of profession, the amount of workload and the degree of discretion matter for how they cope. This paper reads a training cohort with fixed seats as a form of rationing in time: the seats decide how many go now and how many wait.
John D. C. Little, A Proof for the Queuing Formula: L = λW (1961), shows that in a queuing process, under stated conditions, the mean number of units in the system equals λ, where 1/λ is the mean time between two arrivals, times the mean time a unit spends in the system. This paper uses the formula only to name a tie: a number waiting, a rate and a time cannot be set independently, so that a record which fixes the number waiting and the seats has fixed the time only if it has also fixed the rate at which seats come round. Richard C. Larson, OR Forum—Perspectives on Queues: Social Justice and the Psychology of Queueing (1987), holds that attitudes to queues may depend more on other factors than on the wait itself, among them social injustice, defined as violation of first in, first out, and feedback about the likely delay. The paper is speculative by its own account, and this study draws from it only the point that the order of a queue is a question of its own, separate from its length.
The third strand is manpower planning. A. F. Forbes, R. W. Morgan and J. A. Rowntree, Manpower Planning Models in Use in the Civil Service Department (1975), describe the mathematical models a civil service department used for the supply side of manpower planning, how they relate to one another and the problems each suits. This study takes from it only that such models were a family of tools, each suited to its problems; the calculation in these records is a very small member of the family.
4The plan of the paper
Chapter 1 gives the data and method, then reads what the calculation settles (section 1.2) and what the records leave open (section 1.3). An appendix gives the records read, word for word in this study's translation.