Book metadata explained: work, edition, ISBN, and your copy
“Book” can mean four different things. Catalogs become easier to understand once you stop asking one record to represent all of them.
The story is not the edition. The edition is not the ISBN. None of them is quite the copy on your shelf.
WORK: Pride and Prejudice EDITION: Penguin English Library paperback ISBN: identifies that publication / format YOUR COPY: shelf location, condition, notes, loan status
1. The work
The work is the intellectual creation: Jane Austen's Pride and Prejudice, rather than any one printing of it. Searches and reading lists often operate at this level because the distinction between publishers is not always important to the question.
2. The edition or publication
An edition is a particular published version of the work. Publisher, date, language, translator, format, and revision can distinguish one version from another. A translation is not merely a different cover; it changes the text through which you encounter the work.
3. The ISBN
The ISBN is an identifier assigned to a particular publication and product form. Hardcover, paperback, and digital forms can have different ISBNs. The number is a useful lookup key, not a compressed copy of all the metadata around the book.
For the identifier's structure and assignment rules, see what an ISBN identifies.
4. Your physical copy
Your copy is the object you own. It can have an inscription, a missing jacket, annotations, damage, a purchase history, a shelf location, and a loan status. Another person may own a different copy of the same edition and ISBN.
Why apps confuse these layers
Many interfaces show a single card containing title, cover, ISBN, ownership, and reading status. That is convenient, but it can hide the relationships underneath. A search result may describe the right work but the wrong edition. A cover service may attach a neighboring edition's image. Two scans may create two records where you expected one.
What counts as a duplicate?
It depends on the layer. Two records can be duplicates of the same edition, while two physical copies with the same ISBN are not duplicate objects if you genuinely own both. Two different translations belong to the same work but should not be collapsed if the distinction matters to your library.
Translations need more than a title
A translated book still represents the same underlying work, but translator and language are important edition details. If you collect several translations, include those fields prominently rather than relying on the title or cover to distinguish them.
Why cover mismatches happen
Covers belong to particular publications and can also change between printings or markets. Metadata providers may use the best image available rather than a scan of the exact object in your hand. A correct title with the wrong cover can mean the app found the work but attached data from a different edition.
How I would model a home-library catalog
I would keep the model understandable rather than trying to reproduce a research library database. Each owned copy needs a recognizable title and author, useful edition details, identifiers when present, and copy-specific fields such as location, condition, notes, and loans.
The practical list is in what fields to track in a home library catalog. The four-layer model simply explains why those fields do not all describe the same thing.