4 ms·

What’s a good way to group isbns for various versions of the same title? For example, searching Old Man and the Sea on isbndb returns (as you’d expect) many is

What’s a good way to group isbns for various versions of the same title? For example, searching Old Man and the Sea on isbndb returns (as you’d expect) many isbns: https://isbndb.com/search/books/Old%2Bman%2Band%2Bthe%2Bsea https://isbndb.com/search/books/Old%2Bman%2Band%2Bthe%2Bsea Do books have another identifier that logically consolidates editions, foreign language prints, etc.?

cdrini 6y ago
Open Library has a notion of "Works" which group together editions across languages: https://openlibrary.org/works/OL63073W https://openlibrary.org/works/OL63073W (Note the work id is in the url: OL6307W) If you want to get the work id for a given ISBN: https://openlibrary.org/isbn/2070360075.json https://openlibrary.org/isbn/2070360075.json will redirect to the edition page, and there you can get "works[0].key". Or, you can search by the isbn: https://openlibrary.org/search.json?q=isbn:2070360075 https://openlibrary.org/search.json?q=isbn:2070360075
- loosetypes 6y ago
  Thanks. So Open Library implemented that because there's no interoperable, universal concept for literary "works"? And then each of Goodreads, Amazon Retail, Ingram, et al. would likely have their own internal system for isbn identification and grouping accordingly? And I guess that means similarly, there's no isbn-analog for a series or interrelated set of "works"?
  - cdrini 6y ago
    Yep! They do. But the book-universe has largely landed on similar definitions for what a "work" is. E.g. here's GoodReads work id (in the url): https://www.goodreads.com/work/editions/69741-the-old-man-and-the-sea https://www.goodreads.com/work/editions/69741-the-old-man-an... Here's works/editions in schema.org: https://schema.org/CreativeWork https://schema.org/CreativeWork (work) ; https://schema.org/Book https://schema.org/Book (edition) Here's works/editions in Wikidata: https://www.wikidata.org/wiki/Q47461344 https://www.wikidata.org/wiki/Q47461344 (work) ; https://www.wikidata.org/wiki/Q3331189 https://www.wikidata.org/wiki/Q3331189 (edition) (e.g.: https://www.wikidata.org/wiki/Q19320911 https://www.wikidata.org/wiki/Q19320911 ) All of these are loosely based on the FRBR model in librarian science, where a work is a work, and an edition is called an "Expression": https://en.wikipedia.org/wiki/Functional_Requirements_for_Bibliographic_Records https://en.wikipedia.org/wiki/Functional_Requirements_for_Bi... And yeah, there's no isbn-analog for series; each site has its own identifier. Also as a side-note, ISBNs are very inadequate as a unique identifier for an edition; there are frequent cases where an ISBN will be re-used for books with different covers (Dover classics does this a bunch); some modern books published in odd ways are not given an ISBN (e.g. all of these books: https://sorted.club/books/ https://sorted.club/books/ ); and, most importantly, books published before ISBNs went into effect (~1970) have _no way_ of being accessible when using ISBNs as a unique key.
    - loosetypes 6y ago
      Wow, most of that is new to me. I appreciate you taking the time to reply.

cdrini 6y ago

Open Library has a notion of "Works" which group together editions across languages: https://openlibrary.org/works/OL63073W https://openlibrary.org/works/OL63073W (Note the work id is in the url: OL6307W) If you want to get the work id for a given ISBN: https://openlibrary.org/isbn/2070360075.json https://openlibrary.org/isbn/2070360075.json will redirect to the edition page, and there you can get "works[0].key". Or, you can search by the isbn: https://openlibrary.org/search.json?q=isbn:2070360075 https://openlibrary.org/search.json?q=isbn:2070360075

loosetypes 6y ago
Thanks. So Open Library implemented that because there's no interoperable, universal concept for literary "works"? And then each of Goodreads, Amazon Retail, Ingram, et al. would likely have their own internal system for isbn identification and grouping accordingly? And I guess that means similarly, there's no isbn-analog for a series or interrelated set of "works"?
- cdrini 6y ago
  Yep! They do. But the book-universe has largely landed on similar definitions for what a "work" is. E.g. here's GoodReads work id (in the url): https://www.goodreads.com/work/editions/69741-the-old-man-and-the-sea https://www.goodreads.com/work/editions/69741-the-old-man-an... Here's works/editions in schema.org: https://schema.org/CreativeWork https://schema.org/CreativeWork (work) ; https://schema.org/Book https://schema.org/Book (edition) Here's works/editions in Wikidata: https://www.wikidata.org/wiki/Q47461344 https://www.wikidata.org/wiki/Q47461344 (work) ; https://www.wikidata.org/wiki/Q3331189 https://www.wikidata.org/wiki/Q3331189 (edition) (e.g.: https://www.wikidata.org/wiki/Q19320911 https://www.wikidata.org/wiki/Q19320911 ) All of these are loosely based on the FRBR model in librarian science, where a work is a work, and an edition is called an "Expression": https://en.wikipedia.org/wiki/Functional_Requirements_for_Bibliographic_Records https://en.wikipedia.org/wiki/Functional_Requirements_for_Bi... And yeah, there's no isbn-analog for series; each site has its own identifier. Also as a side-note, ISBNs are very inadequate as a unique identifier for an edition; there are frequent cases where an ISBN will be re-used for books with different covers (Dover classics does this a bunch); some modern books published in odd ways are not given an ISBN (e.g. all of these books: https://sorted.club/books/ https://sorted.club/books/ ); and, most importantly, books published before ISBNs went into effect (~1970) have _no way_ of being accessible when using ISBNs as a unique key.
  - loosetypes 6y ago
    Wow, most of that is new to me. I appreciate you taking the time to reply.