How your review is used

Your corrections teach the engine

After each review, @mmcky adds what the review shows to action-translation, as glossary entries, rules and automatic checks. Every later draft in your language uses them. That covers the next lectures in the initial translation, and the sync pull requests that follow it. So a correction you make once should not need to be made again.

For example, a term you correct can become an entry in your language's glossary, so that every later draft uses your term. The next review pull request lists what your review changed, under “New since the last draft”. See Reading a review pull request.

Not every correction becomes a rule. A one-off preference fixes the lecture where you made it, but it may not change the engine. See After your review.

Merged lectures are reference versions

When a reviewed lecture is merged, it becomes the reference version of that lecture. After @mmcky improves the engine, the improved engine translates the lecture again, and the result is compared with your merged version. This checks whether the improvements took effect.

Later versions of the engine are checked against your reference versions too. So each review stays useful after its lecture is merged.

Your review is human judgement

We treat every review as human judgement. That is what makes it a reference for the engine, and it is why reviews are done by hand, with no AI tools. See Review by hand.

Credit

• You are credited as co-author on the commit that applies your suggestions. GitHub shows this on the commit and in the history of the lecture, and the credit stays after the merge.
• Editors and translators are listed in the Translators section of the QuantEcon team page.

Nothing to sign

You do not need to sign up for anything or acknowledge anything for your review to be used in these ways.

For an overview of the whole process, see How an edition is built.