error · Section 5: Procedural writing
Rule 5.3: Write instructions in the imperative form
Write every instruction in a procedure using the imperative (command) verb form, addressed directly to the reader, rather than passive voice or a statement about capability or permission. Do not use "must" immediately before the imperative form unless the instruction is safety-critical or states an important condition.
What a checker looks for
A clear deviation from the rule.
- Flag instructions written in passive voice (e.g. "Oil and grease are to be removed...") instead of the active imperative form (e.g. "Remove oil and grease...").
- Flag instructions phrased as a statement of capability, permission, or fact (e.g. "The test can be continued.") instead of a direct command (e.g. "Continue the test.").
- Flag unnecessary use of "must" immediately before an imperative verb in an instruction that is not safety-critical and does not state a necessary condition.
What this rule does not cover
- "Must" is acceptable immediately before or within an instruction when the instruction is safety-critical (e.g. inside a WARNING) or expresses a necessary condition (e.g. "IF YOU MUST CUT THE WIRE...").
How to fix it
Rewrite the instruction as a direct command beginning with the imperative verb, addressed to the reader. Remove "must" unless the instruction is safety-critical or conditional.
Constraints on the fix
- Preserve the original action and its object exactly; do not change who performs the action.
Examples
Not STE: The test can be continued.
STE: Continue the test.
Not STE: Before you remove the clamp, you must disconnect the hose.
STE: Before you remove the clamp, disconnect the hose.
Related rules
Where this fits in AI governance
A rule a checker can test is a rule a model can be held to. The same discipline, on the governance side of this site.
- Improper output handling
A procedural instruction acted on without a check is the same failure whether the reader is a technician or a downstream system. - Human-in-the-loop review
Procedures are where a generated instruction should not reach a reader without a person having read it first. - Output validation guardrails
A controlled language is a checkable output constraint. This is the control that treats model output as something to be checked rather than trusted.
Source
ASD-STE100 Issue 9, Section 5 - Procedural writing, page 1-5-3.
This page is commentary and does not reproduce the standard. About this rule pack.