Changes to fields and item types for Zotero 5.1

edited March 1, 2017
[edited: These changes will be in Zotero 5.1; see the announcement]

The upcoming release of Zotero 2.1 final is likely to include several (minor) additions and changes to the current set of Zotero item types (e.g. "Book Section") and fields (e.g. "Book Title"). These changes require careful consideration, since they not only affect how bibliographic data is imported, exported and synced, but also how this data can be formatted with Citation Style Language (CSL) styles.

Several members of the Zotero and CSL communities have written up a preliminary list of changes, available at https://github.com/ajlyon/zotero-bits/wiki/Zotero-types-whiteboard. At this point, we would like to call for feedback from Zotero developers and users. Ideally, in the next few weeks, we would like to expand and refine this list so we end up with a collection of well-documented requests that increase the usability of Zotero while minimizing the disruption for existing users.

In considering changes to item types and fields, it is important to remember that item types in Zotero are best understood as a set of fields and a set of ways they tend to be styled. That means that the best argument for new fields or types is that a certain type of data or citation is presently impossible to represent in Zotero. In addition, it is important to note that CSL has its own set of fields (called CSL variables) and item types, and that a mapping is used to make Zotero data accessible to CSL styles. Finally, a small update to CSL 1.0 might be released along Zotero 2.1, which would allow for the addition of new CSL variables.

Comments and contributions can be posted in this thread or in the relevant existing forum threads that discuss changes to item types. You can also request direct access to the wiki on GitHub (see the wiki for details).

[edited for clarity per Rintze's suggestions]
«13456717
  • Two comments:
    - I like the dataset type
    - What is the difference between 'book author' and 'editor' for book sections?

    Maybe more later...

    Generally, I think that it's more important to make two changes (I know, this is kind of unrelated):
    - possibility to define which fields are shown by default whereas the others are only visible after pressing an extend button (the list is to long and it would be nice to have a better overview without the rarely used fields auch as 'Rights' or 'Language')
    - quicker way to copy paste list of authors into item. That always takes a long time when I have to do that (for example for book sections)
  • Greg: This isn't the place for feature requests.
  • What is the difference between 'book author' and 'editor' for book sections?
    Not all book sections are in edited volumes. One typical example would be an introduction written by someone different then the book's author.

    And to clarify - the reason this needs to be discussed with community involvement is that changes to the overall data model are only going to happen very, very rarely and can't really be reverted once they're implemented so we need to get this right. Other features (e.g. the ones requested by Greg) can just be added whenever, because they don't affect the underlying data structure.
  • - I like the dataset type
    Then can you help to fill out the set of fields we'd need for such a type? I really don't know what the type will need.
  • There has been a semi-frequently expressed to desire to make 'document' as generic as possible. That is somewhat better-reflected in ticket 867 than the whiteboard. To make a strong generic type, should document just have every field possible? Similarly, should other "mother types" have all fields of types that seem to be a subset (webpage to blog post would be one example).
  • dataset type - I have been using webpages for that type sofar and I like the idea of not doing it anymore. Attaching stuff such as the survey-instrument or the user-manual is very convenient. I guess important fields are title, creator, distributing institution, abstract, URL, date/collection period, citation required for publications with this dataset, abbreviation...

    I am not sure how the citation required for publications with this dataset should work because sometimes these are books, sometimes, articles. It would be nice to have a dataset type though and when you cite this the correct citation is used.
  • edited December 23, 2010
    Honestly, if we can't nail down meaningful data that ought to be kept on datasets, or if they are treated as books, articles, etc., then there shouldn't be a dataset item type in Zotero, because it follows that datasets are not a coherent category, bibliographically speaking.

    Perhaps you can look at style guides and author guides and compile the rules for datasets in your field's leading journals?

    Edit: Rintze has started work on fleshing out the dataset type proposal: https://github.com/ajlyon/zotero-bits/wiki/DatasetType
    That page also has links to some style guides and previous discussion. Consider what else might be needed to make the type useful.
  • There has been a semi-frequently expressed to desire to make 'document' as generic as possible. That is somewhat better-reflected in ticket 867 than the whiteboard. To make a strong generic type, should document just have every field possible?
    I would like to see this happen.
  • Something I'd like to see is the addition of some new item types. One thing I keep running into is "plays." For example, if I want to add an item for Othello with my lecture notes, video links, articles, etc, I have to add it as something else ("book," perhaps).
  • wpwend - thanks for contributing but that's not yet specific enough to be helpful.
    As per ajlyon's post above:
    In considering changes to item types and fields, it is important to remember that item types in Zotero are best understood as a set of fields and a set of ways they tend to be styled. That means that the best argument for new fields or types is that a certain type of data or citation is presently impossible to represent in Zotero.
    The question is - can you currently cite a play correctly in your preferred citation styles? If yes, we don't need a new item type. If not, why not?
  • Oh, sorry. As far as I can see, you cannot cite them properly. I might be missing something, please point it out if I am, but I don't see a way to export a citation for a play in the correct manner. Please point it out if I am missing something.

    Also, as a matter of "certain type of data," as I tell my students constantly, plays are not books or novels, so I think it would make sense to deferentiate them with a separate item type as a matter of clarity. That might be me just being nit picky, but I think that kind of clarity is important. However, I also understand that may open the floodgates for numerous other things too.
  • Well, how should they be cited?
    The purpose of types in Zotero is not to distinguish between different items analytically, unless that's relevant for citation purposes.
    You can use tags etc. for that purpose.
  • okay, I see what you are saying about dataset type and I agree. There are no guidelines to cite datasets from journals in my discipline I know about. This is done by citing a certain book, article or user-guide...
  • Back to Dataset - using the things Rintze links to, I see three issues that might make a dataset category worthwhile:

    1. The hdl and UNF handlers - so fart those aren't very common - they're a project of Gary King out of Harvard that has, so far, not really caught up - not even in PoliSci, which is Gary's main discipline. In any case - hdl's are a form of URL and should go in that field - I'd add the UNF - so this doesn't really require an item type

    2. Separate info for producer and distributor - APSA requires this for ICPSR (the largest data repository for the social sciences) - and I'd assume that's also the case for sociology
    Eldersveld, Samuel J., John E. Jackson, M. Kent Jennings, Kenneth Lieberthal, Melanie Manion, Michael Oksenberg, Zhefu Chen, Hefeng He, Mingming Shen, Qingkui Xie, Ming Yang, and Fengchun Yang. 1996. Four-County Study of Chinese Local Government and Political Economy, 1990 [computer file] (Study #6805). ICPSR version. Ann Arbor, MI: University of Michigan/Beijing, China: Beijing University [producers], 1994. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor], 1996.
    I have no thoughts on this - but this might require a separate field.

    3. Data Sets are given a different status - in some styles they are treated like monographs with the title in italics (e.g. APA). In some styles (such as the one given on http://thedata.org/citation/standard ) they are treated as articles and put in quotation marks. In the third set of styles (such as the APSA style) they are treated as neither and given in normal font without quotation marks, although monographs are in italics and articles in quotation marks otherwise.

    At least in the disciplines I know - PoliSci, Sociology, and Economics - I have never seen any of the other requested information on the various data set threads (such as sampling group etc.) in any citation.
    Imho so far this doesn't look like a separate data set type is needed. It would be good to hear from some other disciplines, though - I'm thinking geography, psych, maybe some of the statistically oriented life-sciences like epidemiology etc.
  • The hdl and UNF are respectively a stable URI (which in the example at http://thedata.org/citation/standard is not a URL, but more like a DOI) and a hash value. Both seem useful.
  • New issue:
    I would like to see a "type" variable included in book. That would help e.g. for styles that want things like "ebook" "digital image" or even "kindle book" listed. I'll dig up the links when I get the chance - erazlogo once said she didn't like the idea, so I wanted to open that for discussion rather than just adding it to the list on github
  • edited December 23, 2010
    "Format", mapped to the CSL variable "medium", could perhaps work for that as well.
  • ah yes, I had forgotten about the medium variable - that would be better, I think.
  • @wpwend42: Since I told you on Twitter to post here, I ought to be more clear. Zotero's data model is built around ways we record metadata about publications and other produced or performed works. I'm not sure what problems you currently are having as you try to use Zotero to work effectively with plays, but I can imagine some possibilities. It is, for example, not very straightforward (or not possible?) to represent the play Hamlet in Zotero. There are ways to represent printed editions of it (manuscript, book, book section) and ways to represent specific perfomances of it (presentation, video recording), but the underlying creative work doesn't really fit into this model.

    Of course, it's obvious to any theater scholar that there isn't a single underlying Hamlet-- we can only talk about print, stage, screen manifestations. This is something that will some day be addressed by the long-discussed hierarchical item model, but it's a hard problem theoretically and practically.

    For now, we have a window of opportunity to make changes to the specific fields and types at our disposal -- to see if plays need something Zotero can't do, just explain what data you need to keep on plays and what citations look like.

    We'll try to figure out whether the current model can handle it and propose changes if it can't. The current system is more flexible than it looks.
  • @Rintze & adamsmith: I agree that Format should be added to Book/Book Section -- but I worry that we might need to add it to even more formats. I went ahead and added it to the whiteboard.
  • edited December 24, 2010
    It appears that the book review discussion had been overlooked in preparing the proposal. Fortunately, the thread has been revived. In that thread, erazlogo wrote:
    The easiest way to add book reviews is to add "reviewed title" field and "reviewed author"/"reviewed editor" author types to all article item types--this allows for correct bibliographic formatting for journal/magazine/newspaper books reviews.
    That sounds good to me.
  • That would require changes to CSL as well, as there is currently nothing to map "reviewed title" and "reviewed author"/"reviewed editor" to.
  • So reviewed author is currently just ignored?
  • Frank's mapping overview ( http://gsl-nagoya-u.net/http/pub/csl-fields/index.html ) doesn't include creator roles, so I'm not sure.
  • I can imagine some possibilities. It is, for example, not very straightforward (or not possible?) to represent the play Hamlet in Zotero. There are ways to represent printed editions of it (manuscript, book, book section) and ways to represent specific perfomances of it (presentation, video recording), but the underlying creative work doesn't really fit into this model.

    This is more or less what I am trying to get at here. Thanks!
  • So reviewed author is currently just ignored?
    It turns out that reviewed author is used in Bibliontology RDF export, import from the Bryn Mawr Classical Review, and that's all. It is item type "reviewedAuthor" (id 27), but it doesn't look like Zotero maps it to anything for CSL, unlessed I've missed something in my code-diving.
  • edited December 31, 2010
    A current limitation of Zotero/CSL is their inability to graciously handle label-type styles. CSL contains a "citation-label" variable, but it isn't accessible through Zotero. Instead, some people use the "Extra"/*extra* field that maps to *note*: http://forums.zotero.org/discussion/5239/first-letter-of-author-as-citation/

    Would it make sense to add a "Citation Label" field to all item types, or would this clutter the UI too much?
  • I think that this discussion has pretty much ended; if anyone has additional proposals that are not reflected on the whiteboard, now would be the time to add them. Otherwise, it's probably time to clean up the present proposals and make a final list of specific changes.

    As we do so, we still need clarification on:
    * Artwork. The whiteboard has no concrete proposals on what can or should be changed with that type. Without some clarification and justification, it's not going to happen for 2.1.
    * Music Score. Older discussions acknowledges the need, but we have no details on necessary fields.
    * Exhibition Catalogue
    * Series Titles and Series Text. These fields are unclear to begin with-- the whiteboard includes a proposal to repurpose one as "section"; see the whiteboard.

    Between these changes, standalone, and multilingual, the 2.1 release of Zotero is going to be a very exciting one.
  • @Rintze: I'm leaning in favor of supporting Citation Label, but I'd like to see field hiding as well, since it (and the present proposal in general) threaten to really clog up the metadata pane, especially for small-screen users like me.
  • @ajlyon: agreed that field hiding would be very nice. There will be other fields that many users won't use, like the proposed original-date/original-publisher/original-publisher-place fields.
Sign In or Register to comment.