Rules historically came in two forms: thick rules riddled with examples, caveats, and room for discretion, and thin rules stripped to explicit, exhaustive formulations that leave nothing to judgment. The thin, algorithmic rule is a relatively recent achievement, not the natural or original form of a rule.
Thin, exception-free rules only work when the surrounding world is predictable and well-ordered; in variable or unstable conditions, rules must stay thick and flexible to remain usable. The drive toward rigid specification therefore tends to accompany efforts to standardize and stabilize the environment itself.
For much of Western history the paradigm of a rule was a model or paradigm to be imitated, such as a saint's life or a master craftsman's example, rather than a set of written-out directions. Following a rule meant cultivating judgment by emulation, not mechanically executing steps.
Benedictine and later monastic rules were written loosely and expected the abbot to adapt them to persons and circumstances, treating written rules as guidance requiring interpretation rather than literal commands. The discretion granted to the enforcer was a feature of the rule, not a failure of it.
Historically a rule and its exceptions were understood together: knowing when not to apply a rule was part of knowing the rule, and the phrase 'the exception proves the rule' originally meant the exception tests and defines the rule's proper scope. Viewing exceptions as defects to be eliminated is a later attitude that changes what a rule is taken to be.
Algorithms as mechanical, step-by-step procedures requiring no insight existed long before computers, in calculation rules and bureaucratic routines, and their appeal lay in removing the need for individual judgment. The historical move toward algorithmic rules was driven by distrust of discretion and a desire for uniformity across many hands.