longhotsummer
commented
4 years ago A relatively simple workaround for the time being is that reviewers should highlight more text or adjacent text. For structural changes, they should highlight text higher up in the hierarchy, e.g. the subsection number if paragraphs need to change.
This should definitely be encouraged.
It's very difficult (impossible?) to work out what to do with annotations that were on text that has subsequently been deleted or changed heavily. The more text that is highlighted, the more context is available to work out where the annotation should go.
If an entire part of the hierarchy is removed and the annotation was on that part, it's equally difficult to know what to do with that annotation.
A long-term compromise could be to show "floating" annotations as close to their original source as possible, when the annotation cannot be bound to its original context.
Feedback from a reviewer: if text is highlighted that should be deleted, when the document is resubmitted the annotation has moved to a different instance of that word.
At second review, a reviewer should be able to just work through the comments they left in the first review, but this isn't feasible if the comments move around or disappear.
A relatively simple workaround for the time being is that reviewers should highlight more text or adjacent text. For structural changes, they should highlight text higher up in the hierarchy, e.g. the subsection number if paragraphs need to change.
Until the annotating experience is full-proof, which won't be in the short term, these workarounds should be communicated clearly to all reviewers.