Conversation

Notices

1. Embed this notice
   Eaton (eaton@phire.place)'s status on Tuesday, 11-Jun-2024 23:16:51 JST Eaton

   One of the signs you're Not Well is that putting a 'reading list' section on your blog starts with building a rotating-proxy amazon scraper and a NLP powered metadata scrubbing engine that fixes common problems — Edition names jammed onto the end of Publisher names, publication dates appended to Imprint names, subtitles that are actually author lists, etc.

   But, like… what am I supposed to do? Let BAD DATA just STAY BAD?

   *wild-eyed, twitchy stare*

   • Embed this notice
     Eaton (eaton@phire.place)'s status on Tuesday, 11-Jun-2024 23:21:41 JST Eaton

     in reply to

     Structured book metadata is less of a contract and more of a puzzle game, in which you try to figure out whether ‘PENGUIN' appearing in the Publisher, Format, and Edition fields of a book about penguins is a sign that Penguin/Random House published it, that the metadata is duplicated, or that someone’s enthusiastic 5 year old got ahold of the keyboard on release day.

   • Embed this notice
     Mark Llobrera (markllobrera@mastodon.social)'s status on Wednesday, 12-Jun-2024 01:32:03 JST Mark Llobrera

     in reply to

     @eaton It’s such a headache with my own reading log, and I view any API-pulled data as a mere starting point that needs human intervention

   • Embed this notice
     Eaton (eaton@phire.place)'s status on Wednesday, 12-Jun-2024 01:32:03 JST Eaton

     in reply to

     • Mark Llobrera

     @markllobrera 100%. I've got things in ... DECENT shape, though not so good that I’d be comfortable doing (say) auto-generation of author-name index pages.