internetarchive / openlibrary

One webpage for every book ever published!
https://openlibrary.org
GNU Affero General Public License v3.0
5.11k stars 1.34k forks source link

ImportBot is creating authors based on faulty AMZ record rather than accurate WorldCat entry #4065

Open LeadSongDog opened 3 years ago

LeadSongDog commented 3 years ago

Evidence / Screenshot (if possible)

Relevant url?

https://openlibrary.org/works/OL23083613W/Je_couds_pour_bébé?m=history

Steps to Reproduce

  1. Go to https://openlibrary.org/works/OL23083613W
  2. Click Amazon link, note wrong author "Sonia Roy"
  3. Go back.
  4. Click Worldcat link, note correct author "Isabelle Leloup", as shown on cover.

Details

Proposal & Constraints

Stop bot from trusting AMZ (or any one bookseller) as a sole reference for authority. Check for that ASIN or ISBN in any federated catalog (Worldcat, KVK, etc)

e.g. KVK finds that ISBN in the NLF and Worldcat, both catalogues give the correct author, not Sonia Roy (who is listed, but only as a collaborator):

https://kvk.bibliothek.kit.edu/hylib-bin/kvk/nph-kvk2.cgi?maske=kvk-redesign&lang=en&title=KIT-Bibliothek%3A+Karlsruher+Virtueller+Katalog+KVK+%3A+Ergebnisanzeige&head=%2F%2Fkvk.bibliothek.kit.edu%2Fasset%2Fhtml%2Fhead.html&header=%2F%2Fkvk.bibliothek.kit.edu%2Fasset%2Fhtml%2Fheader.html&spacer=%2F%2Fkvk.bibliothek.kit.edu%2Fasset%2Fhtml%2Fspacer.html&footer=%2F%2Fkvk.bibliothek.kit.edu%2Fasset%2Fhtml%2Ffooter.html&css=none&input-charset=utf-8&ALL=&TI=&AU=&CI=&ST=&PY=&SB=9782848317540&SS=&PU=&kataloge=NLAU&kataloge=VERBUND_BELGIEN&kataloge=DAENEMARK_REX&kataloge=EROMM&kataloge=ESTER&kataloge=NB_FINNLAND&kataloge=FINNLAND_VERBUND&kataloge=BNF_PARIS&kataloge=ABES&kataloge=COPAC&kataloge=BL&kataloge=NB_ISRAEL&kataloge=VERBUND_ISRAEL&kataloge=EDIT16&kataloge=ITALIEN_VERBUND&kataloge=ITALIEN_SERIALS&kataloge=CISTI&kataloge=NLCA&kataloge=LETTLAND_VERBUND&kataloge=LUXEMBURG&kataloge=NB_NIEDERLANDE&kataloge=VERBUND_NORWEGEN&kataloge=NB_POLEN&kataloge=VERBUND_POLEN&kataloge=PORTUGAL&kataloge=STAATSBIB_RUSSLAND&kataloge=VERBUND_SCHWEDEN&kataloge=BNE&kataloge=REBIUN&kataloge=NB_TSCHECHIEN&kataloge=NB_UNGARN&kataloge=NLM&kataloge=WORLDCAT&ref=direct&client-js=yes

Related files

Stakeholders

LeadSongDog commented 3 years ago

@hornc the bot is still doing this, please stop it. We should not have to clean up crap like: https://openlibrary.org/books/OL32785017M/Brooklyn (created at https://openlibrary.org/recentchanges/2021/07/12/add-book/83586306 ) or https://openlibrary.org/books/OL32785034M/Kalligrafie AMZ only pseudo-books with bogus authors

hornc commented 3 years ago

@mekarpeles , this is one for you. @LeadSongDog unfortunately I do not control the behaviour of import bot and it's over-reliance on Amazon data. I have been able to import the latest publicly available LOC MARC records into OL, and other MARC records, but the Amazon imports are a separate process.

LeadSongDog commented 3 years ago

@hornc Thank you for clarifying that. @mekarpeles The First Law of Holes pertains.