Deterministic and Feasible (D-F)
Task nature where a correct answer can be computed exactly at practical scale
Deterministic and feasible (D-F) is one of three task-nature classes in the When the Commitment Boundary Needs Reinforcing calibration: a correct answer exists, the same inputs always yield that answer, and computing it exactly is practical at the scale and latency the organisation needs.
Examples: tax calculation, sorting, rules-based eligibility, string matching, train position from speed × time given current measurements.
D-F is not about simplicity. Sorting a million records is complex and still D-F. “Is this email spam?” sounds simple but is usually genuinely probabilistic (P) — judgment under ambiguous evidence.
At execution: once a correct algorithm is in place, each instance is solved by logical inference from inputs — not by human or AI judgment at the commitment boundary. That is Rule 1 (design — finding or improving the algorithm — is a separate, intelligent act; see the article).
Primary mechanism: deterministic algorithm, testing proportionate to outcome severity. When many are affected: publish evidence and derivation — not average opinions (When Trustworthiness Is Not Enough).
Sibling classes: D-I and P. Matrix: Decision Framework map.