Zotero usage scenarios (word processors, Mac...)

Zotero has shaken things up for me. It's an exciting development and has already helped my research immensely, but I can't exactly figure out how I want to use it in my writing.

I used to manage my citations with Bookends and write in Mellel. But after switching to Zotero I went back to Word and used the plug-in. This is a pretty sweet set up, but I was thinking of just using the drag and drop from Zotero into Scrivener (Chicago Style footnotes), since I like Scrivener much better for big projects--you lose the automation, but this approach also keeps it simple, no?

I work in visual studies and literature, so I probably cite fewer sources than many others (like historians...) This is also a consideration.

In any case, I was wondering how others use Zotero--apart from the plug in--to write articles and such. Any shortcuts? Or is it a bad idea not to use the plug in?

Any suggestions or ideas would be helpful.

thanks.
  • Hmm ... good questions, and one's I've struggled with myself.

    I'm the author of the citation style language Zotero uses, and I've always said that my goal in creating and promoting CSL is to help create the conditions where none of us have to think about citations much at all.

    We're a long way from being there, in part because formatting is still tied to specific bibliographic applications, which in turn choose to support specific editing solutions, each in somewhat different ways.

    The consequence is a general lack of choice. I personally don't like writing in word processors; I find they slow me down and get in the way of thinking through ideas (and I also like using revision control systems to manage the evolution of the documents, handle syncing between machines, etc.). I imagine my perspective on this is the same as your's when you say you prefer using Scrivener.

    So ATM you're left with a difficult choice: use Word (or OpenOffice) and get automatic citations, or use something else and maintain them all manually.

    To me this is an unacceptable state of affairs, and I'm trying where I can to enhance citation support in other contexts as well; for example, in markup languages like markdown. Ultimately, we need real citation support across the range of common authoring solutions, including what will become increasingly the norm: web-based solutions like Google Docs.

    I'd say if you really prefer the "other" option to Word/OOo, just go with that. You could also always look into ways to encode citation logic somehow yourself, maybe by adding labels to your other field, and adding those to your documents?

    Sorry: there really aren't any good answers to your question ATM.
  • Hmm, I wonder if Zotero (or really CSL) needs a generic post-processor that can turn some kind of mini-markup {{Smith, 2000, AFewAmbiguity-ResolvingTitleWords}} into real citations. RefWorks does postprocessing like this for RTF files, AFAIK, but it would be nice if it were smart and didn't absolutely require a "citation key" or unique database ID. Then it wouldn't be dependant on having a database-record-inserter (like the present Word/OOo macros or RefWorks' WriteNCite) which could interact with your writing software. Therefore it would make the editing environment irrelevant. I'm sure this is not true for everyone, but with my database, I think I could make a thousand citations before "Surname, publication year and a few significant title words" would give a bad result if the parser were intelligent. And the danger of false results could be further minimized if the postprocessor gave relevant warnings for ambiguities. Such a post processor should probably also work with a proper citation key (using global URIs, for sharablity), but it would be nice for the user with more modest requirements if something like the above suggestion would also work. (It would actually look good in Markdown, for one thing, though I suppose non-inline markup could be made to work as well.)


    Ideally such an engine would allow (1) some flexibility in specifying markup, to avoid clashes with the markup languges involoved and (2) a minimal set of citation transformations (omit Surname, Title, date, insert ). If it also (3) worked with RTF HTML, ODF (for the non-OOo ODF programs), Markdown/Pandoc, or Markdown/Multi-markdown, and, say, ReST out of the box, it would be usable for a whopping lot of environments, with the ability to add more as needed.
  • edited April 8, 2008
    The more I think about this, the more I like it. There are two parts to the idea:

    (1) A post-processor which would read documents in an extensible number of input formats. It would use your local Zotero database to turn citation references (presumably written as plain text with lightweight markup) into bibliographic citations with proper formatting using any CSL style.

    (2) Such a post-processor could be made to be both robust and human-friendly.

    (a) Robust: It could use universal item IDs to reference items to be cited. That way, documents would be portable among users, and usable even by people who don't have the appropriate item in their local database. (The parser would be configured to use web sources as needed to get bibliographic information.) When writing, you could insert your citations with the help of your local Zotero installation, a browser-based webapp or small desktop client.

    (b) Friendly: It could also be configured to work with less robust, but writer-friendly citations like {{Smith, 2000, Interstellar}} or {{Smith, 2000}} or {{Smith, Interstellar}} or {{Smith, ISBN:1234123456}}. This would let a person compose citations while writing regardless of the state of connectivity of the computer, or the accessibility of bibliographic information. It would keep the human overhead low during the act of writing, which is a Kind Thing for Users. A smart post-processor would look for bibliographic data in the local Zotero database at the time of parsing, flag ambiguities, and perhaps go shopping on the web (with optional user babysitting) when the information is not found. It should also have the option to convert 'friendly' to robust citations.

    Such a post-processor could have a plugin architecture, for extensibility. In the hands of an ambitious coder it could become a CSL replacement for BibTeX, for example.

    Problems it could help solve:

    (a) Automatic citation for all, even the unwashed users of funky markup languages, beta version webapps, and lightweight cross-platform word processors. Any documented file format is a candidate for compatibility. (Even, say, upstart ISO-contender office formats.)
    (b) Interoperability. Users of SomeNewWebapp could roundtrip documents with users of the latest desktop word processors, assuming file format conversion by the word processors themselves.
    (c) With 'friendly' citations, I can cite an item with reasonable reliability even before I've had a chance to add it to my database. I often want to do this, and don't want to interrupt my writing to do it.
    (d) With 'friendly' citations I can cite no matter what sort of platform I happen to be typing on at the moment, what the state of its connectivity, or where my personal Zotero database is, as long as I have access to a few distinguishing details.

    Problems it doesn't solve:
    (a) Robust item IDs won't be pretty, I assume. (However some desktop word processors could be trained to hide them, and lightweight markup languages can possibly use non-inline mechanisms to keep the ugly bits out of text paragraphs).
    (b) They aren't WYSIWYG. That would have to be accomplished by another means, presumably involving the cooperation of every WP/editor you hope to work with.
    (b) At the moment, we don't have universal item IDs anyway, except perhaps for published books.
    (c) 'Friendly' citations are by definition not robust.
    (d) Plenty more, I'm sure.

    Is something like this the way to go? (Bruce?) Is it doable from a coding standpoint?
  • edited April 8, 2008
    Yes, Scot, you've got the idea.

    Really what we need is an API for citations, and then different input/output drivers for different applications and formats.

    So, for example, you have a generic notion of a citation and a reference. A citation is an ordered list of references. Each reference contains, say, a local id label, and a global id (a URI), plus optional locators, prefix and suffix, etc.

    You then write, let's say, a markdown input filter that maps that format to the API. For ODF, you use its new text:meta-field. For OOXML, use their citation field. Adding a new format is just a question of adding a new driver.

    At the OpenOffice bib project, we had worked on this awhile ago, but as I think about it, it really ought to be a generic problem, and that the experiences of Zotero ought to inform it. Such an API ought to be possible to implement as a system service, BTW (certainly in OS X).
  • A citations API as a system service would be lovely.

    In the meantime I figured out that using Exposé can mimic some of the functionality one misses by not using Open or Word with Zotero. In one Expose section is my WP and in the next I place Firefox--a key stroke can take me back and forth easily... A little bit like the way you can call Zotero from Word using the citation button. A very little bit. But I quite like the way it's working out for me (and it allows me to write in Scrivener and Nisus which I rather prefer).

    best,
    s
  • I am now also very much in favor of the generic approach as proposed by bdarcus! Instead of developing many plugins for many different versions of many different programs, on many different platforms, we should just have
    (1) a simple (text-only?) and reliable format for reference-marks which would have to be inserted into the documents (it would suffice if that would be done by selecting the reference in Zotero, applying a certain shortcut and pasting the mark into the document);
    (2) a reliable cross-platform post-processor which would process the final document and change the reference marks into full bibliographic data.
    The post-processor should be able to process the most important formats (e.g. rtf, odf, doc), but it would not have to deal with all kinds of APIs of many different programs. Currently, formatting bibiliographies is one of the (few) clear weaknesses of Zotero in comparison with programs like EndNote. However, if we abstain from imitating nice but, at least in my view, completely unnecessary features (like the 'cite while you write' functionality), and focus instead on reliability and flexibility (in terms of cross-platform compatibility and a simple and well-documented front-end for modifying bibliographic styles), Zotero could become superior in this area as well.
  • No need to cross-post. Zotero 2 has rtfscan.
  • One potential of such an API is the ability for regular citations to become akin to a hyperlink that points the material cited through some kind of mediating layer.

    If I have a PDF of some article in Zotero and I cite a particular page, I ought to be able to e.g. Control-Click it and have the document open to that page. Then if I send my document to someone else and they don't have the paper and they follow the citation - Control-Click it or whatever, they're directed to some intermediary that can do something with the information. Give them additional reference data not present in my citation, do a journal search for that article, etc.

    You could also use it to do things like automatically verify the accuracy of quotations. If I'm quoting and citing a paper in the same line, it should be obvious from syntax what I'm doing. If the document, page number, and quoted text can be extruded from that, then if I were to accidentally misquote someone then that should be evident.

    I suppose you developers have already thought that through, but it's just dawning on me what a difference standards in this area could make to scholarship.
  • In response to samuelas' good question ("I was wondering how others use Zotero--apart from the plug in--to write articles and such"), I started out thinking I would use the copy-and-paste output technique for articles that require the Chicago style in footnotes. This turned out to be a major mistake. When I edited my manuscript and altered the sequence of the notes, my edits made a mess of the retrospective references (ibid. vs. short titles). When I add citations with the plug-in, I need only choose the Refresh option in Zotero's menu after editing in order to repair the damage.

    This problem doesn't exist with in-text author-date citations, of course, but I still find myself using the plug-in. It offers the option of suppressing the author's name, as in the following:
    According to Latour, dualities such as technology vs. society and nature vs. culture are the consequence, not the cause, of sociotechnical activity (1992).
    I also really like the handling of multiple citation and the option of hand-editing the citation for special cases.

    To put a bibliography together quickly (e.g., for a student), the copy-and-paste output technique comes in handy.
Sign In or Register to comment.