internetarchive
Removed covers should not be auto-reimported
Cover imports are weirdly aggressive. If I open a coverless edition, frequently a cover is instantly imported. If I remove that cover and add an ocaid (which has a real cover), the deleted cover is reimported and I have to remove it again.
The cover imports do not appear as cover edits in the history.
Evidence / Screenshot (if possible)
Relevant url?
https://openlibrary.org/books/OL46535237M/Merry_Christmas_Alex_Cross?m=history
Steps to Reproduce
- Go to ...an edition with no cover.
- Do ...Click edit to see if a cover is auto-imported. If so, remove the cover and save it. See if the cover was immediately re-imported on save.
- Actual: Removed cover was immediately reimported on saving the new edits.
- Expected: Removed cover should not be reimported. In this particular case, since I added an ocaid, importing the item's actual cover would have been appropriate and welcome.
Details
- Logged in (Y/N)?
- Browser type/version?
- Operating system?
- Environment (prod/dev/local)? prod
Proposal & Constraints
Covers imported from a retailer are frequently be incorrect or low-quality. If a human editor removes the imported cover, it should not be reimported.
Ideally, when an item has an ocaid added, it would be useful if the cover of the actual item were imported (rather than a stock image from a retailer).
Related files
Stakeholders
@mekarpeles
That import has much bigger issues than just a wrong cover:
- the book was scanned in Aug 2020, but apparently never had an OpenLibrary record created
- the book has real librarian-created metadata available, but instead the extremely limited and poor quality Amazon metadata was used
- Due to 2, it didn't match to the correct author record (because the birthdate from the MARC record wasn't used)
- Because the author is wrong, there's no chance for the edition to be matched with the correct work
- Due to 2, all the subjects from the MARC record are missing, so the new work isn't findable by any of them
- Due to 2, the fact that it's a large print edition isn't recorded
In the last year or so there's been an absolute explosion in poor quality imports leading to cascading data quality issues like this, inundating the poor volunteers who are shoveling against the machine driven tide of slop.
