Skip to content

docs: explain what each measured attribute actually means #67

Description

@jothimani-rajendran

A statement carries rows for reachability, conformance, decision, efficacy,
fault.malformedOutput, fault.scriptMissing, fault.interpreterMissing, fault.timeout,
interference, and the two cost.* rows. A reader meeting a statement for the first time can see
the verdicts without knowing what was actually asked.

What to do: write one short section per attribute — what question it asks, what the producer
does to answer it, and what each verdict means. Concrete: name the actual probe, not a paraphrase
of the attribute's name.

The part that matters most: be explicit about when each attribute reports absence rather
than a bad result, and why those are different. NotApplicable on an artifact that declares no
hooks is not a failure; treating it as one is exactly the misreading this format exists to
prevent. The same distinction is why the row-basis figure draws "not measured this run" in neutral
grey rather than at the pale end of a ramp.

Method: read a real statement under paper/measurements/ alongside the code that produces each
row, and describe what the code does. An attribute's name is a label, not a specification, and
several of them are easy to guess wrong from the name alone.

Sized to an afternoon. Files: docs/.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions