Aglet

Learn from recurring document format drift

Repeated format findings are easier to resolve when binary changes are separated from semantic changes. Learning should preserve marked source and output pairs, document the properties that matter to readers, and make intentional version or option changes reviewable by downstream owners.

Keep the lesson for the next incident

  1. Cluster output changes

    Group findings by extension, metadata, text, page structure, rendering, compression, option, version, and nondeterministic behavior. Record source marker, processing context, changed property, and downstream consequence.

  2. Write the stability contract

    State which output properties must remain stable, which may vary, and what context or version explains an intentional change. Include paired artifacts with text and visual landmarks plus the expected filename and format.

  3. Trigger comparison review

    Require the paired fixture when changing processors, formats, metadata handling, options, versions, or downstream handoffs. Assign an owner for compatibility and record how a reviewer distinguishes harmless bytes from changed document meaning.

What to carry forward

The learning record is complete when output stability is defined by semantic properties, paired artifacts show accepted variation, and context changes are explicit. If a consumer requires a stricter contract, record that dependency rather than generalizing one output rule to every use.

Technical background: MDN reference.

Keep the decision with the work.

Use a Work Item in Aglet to record the problem, the evidence you have, and the next decision. Add an owner and priority, then keep updates in the discussion so the next person can follow the reasoning.

Create an account See the product workflow