The paper offers a new evaluation criterion for topic models, based on a precision/recall trade off. This criterion is operationalized in a new implementation. Topic modeling is a well-studied field, so finding promising new directions is exciting. The reviewers found the author response compelling and are all in favor of acceptance.