Creating book chapters creates duplicates because of the DOI

q
Zotero 8 seems to consider the DOI for identifying duplicates.

If I create a book that has a DOI, and then choose "create book section" (right click), the DOI is maintained. As a result, all book sections of the same book are considered duplicates in the "Duplicate Items" view.

My inkling is that the DOI should not be inherited by the book chapters. Chapters can have their own DOI.
  • dstillman Zotero Team
    Yeah, removing the DOI is probably right here. We'll create a ticket. Thanks.
  • dstillman Zotero Team
    Fixed now in 8.0.2
  • jschneider
    I'm on 8.0.3, and I see a number of book chapters considered duplicates. Can you please reopen the issue? Perhaps other identifiers (ISBN) are being used as well on book chapters?

    Here are some examples.
    https://s3.amazonaws.com/zotero.org/images/forums/u13493/fs535y4gpztfc5cn82rr.png
    https://s3.amazonaws.com/zotero.org/images/forums/u13493/z73nl9wq03xe6pl35u18.png
    https://s3.amazonaws.com/zotero.org/images/forums/u13493/kwoul9h7sxbopkpjjuy3.png
  • dstillman Zotero Team
    You'd have to show us the data for each one.
  • jschneider
    I sent a Zotero RDF file to support.
  • dstillman Zotero Team
    edited 3 days ago
    I really just meant a screenshot of each item’s metadata pane.
  • q
    @jschneider, 8.0.2 fixed the issue so that the function of creating book sections no longer inherits DOI. If the book sections actually have the same DOI (this is what the screenshot with the metadata would show), I think this is correct behaviour.
  • dstillman Zotero Team
    edited 3 days ago
    Yes, if you created book sections with duplicate DOIs between when the new fields were added and 8.0.2, you'd have to remove the DOIs manually.
  • bwiernik
    I actually disagree about removing the DOI. For APA style at least, the expectation would be to cite the DOI for the whole book when citing a chapter of the chapters don’t have their own DOI. I think it would be better to update the duplicate detection logic to not count as duplicate if the DOI matches and the items are books/chapters with different titles
  • adamsmith
    Isn't it pretty rare for edited books to have DOIs if their chapters don't? I don't think I have seen that
  • bwiernik
    I think it varies a lot by publisher—about half of the book chapters I’ve written don’t have their own DOI. I’ll check which publishers
  • aborel
    I don't know why, but it seems to be the standard practice of at least one publisher (Routledge) to not assign a DOI to its book chapters, only to the ebook itself.
  • q
    @aborel, I've just checked on the Routledge page for an example:

    Book: https://www.taylorfrancis.com/books/edit/10.4324/9781003458418/routledge-companion-east-asian-historiography-edward-wang (has DOI: https://doi.org/10.4324/9781003458418)

    Individual chapter: e.g. https://www.taylorfrancis.com/chapters/edit/10.4324/9781003458418-5/writing-histories-tang-618–907-liu-zhiji-critical-historiography-victor-cunrui-xiong (doesn't seem to have a DOI).

    In my reading, the DOI refers to the book, and not the individual chapters. An ISBN by definition refers to the book. In this sense, the behaviour in 8.0.3 seems intuitive to me, but I'm not an expert on these things.
  • poettli
    edited 2 days ago
    Individual chapter: e.g. https://www.taylorfrancis.com/chapters/edit/10.4324/9781003458418-5/writing-histories-tang-618–907-liu-zhiji-critical-historiography-victor-cunrui-xiong (doesn't seem to have a DOI).
    10.4324/9781003458418-5 is the DOI of this chapter. (You can check by clicking https://doi.org/10.4324/9781003458418-5)
  • aborel
    Good, it sounds like they eventually understood the need for more granularity! It's not really obvious, and the current translator doesn't seem to capture the DOI, but since it is available in the URL and somewhere in the HTML we could improve that.
  • bwiernik
    Yeah, the examples I found in my library are from Elgar or Routlege as well. Both have started to register individual chapter DOIs too. Elgar seems to be consistent there—didn't find any examples in a quick review where the chapters didn't yet have their own DOIs. Routlege is still sporadic:

    Some examples:

    Whole book DOI registered:
    https://doi.org/10.4324/9780203111505
    https://doi.org/10.4324/9780203809808


    But individual chapter DOIs are not (despite being used for internal URLs in the publisher catalog).
    https://doi.org/10.4324/9780203111505-5
    https://doi.org/10.4324/9780203809808-23
  • bwiernik
    Given that it seems that it is now quite rare for chapters to not have their own DOIs, I've changed my mind and think removing the the DOI on "Create Book Section" is probably the right call.
  • iansethmcphail
    There are indeed a number of book chapters in my collection that I can't seem to find the DOIs for (all seem to be from Routledge).

    Not sure what the best option is here. My understanding is that at least in APA, I'd want to have the DOI for the entire book if the chapter doesn't have its own. However, this indeed clogs up the duplicates in Zotero.
  • jschneider
    The reality is that DOIs are at different levels of granularity. (Medical conference abstracts are my current favorite example - old ones typically have identifiers at the PROCEEDINGS level - new ones may have identifiers at the abstract level - but the identifiers still get you "close" - depends on what you want identifiers for, access vs. VERY strict identification.)

    @q: That is worse and it does nothing for my benefit (since I DO want book chapters to have the book-level DOI). Filtering the duplicates is what would ACTUALLY help me. A related problem is that there are often duplicates I *have* reviewed (e.g., preprint has an extended version of a publication so I want to keep both even though they are related, they are not mergeable). If you want to talk more about this let me know.

  • adamsmith
    Yeah, there's no perfect solution here but since more chapters have DOIs than not, for the item creation I think the current behavior in 8.03, which is that Create Book Section removes the DOI but Create Duplicate keeps it (thus leaving a possible workaround for jschneider's use case) seems right.

    In parallel, I think the duplicate detection needs to change. Since there clearly are examples of chapters without DOIs in books with (or abstracts in proceedings -- same logic), relying on DOIs as sufficient indicators of duplicate items won't work anymore (I generally think duplicate detection could use some love -- fine tuning, mark non duplicate, merge all duplicates -- there's tons of stuff people want that currently requires plugins but seems like it should mostly be built in).
Sign In or Register to comment.