Library / Systems Engineering Principles Framework
Jump to passage
In this reading

Link to current text

Published source confirmed at last check

Source changed 2026-10-03 02:22:15 UTC · snapshot created 2026-10-03 03:38:22 UTC · last check 2026-10-03 04:10:10 UTC

SYSE.46:4.1 - Select the claim and comparison

State the intended task family, actual result, protected effects, available supports and proposed reliance. Recover the actual performing arrangement. For a technical agent this includes model/settings, instructions, next-input assembly, memory, tools/editions, executor and environment. For a person it includes applicable prior performance, instructions, available aids, working material and conditions such as task order or fatigue. Identify what changed, what can be held fixed and what must instead be recorded.

Choose a baseline that represents a real alternative, including a direct deterministic operation or human-supported arrangement when appropriate. Compare the same receiving result and consequential conditions. Separate result quality, latency, cost, necessary access, restraint and side effects; combine them only when the decision supplies a justified trade-off.

Choose observations from failures that could change the reliance. A model-output test cannot establish a state-changing result. Obtain the domain’s acceptance predicates and the evidence that connects the measured signal to them.