please add better integration with Scrivener

I have tried the rtf scan, and it doesn't work using the latest versions of Zotero, Scrivener, and Snow Leopard. It would be great to have a plug-in just as there is for Word and Open Office. Thanks.
«134
  • it's pretty clear that Zotero isn't going to provide that. It's likely that the RTF scan will improve, but if there is going to be a Zotero plugin for Scrivener it'll have to come from Scrivener or a third party.
  • edited May 20, 2011
    How do you know this? Have others tried and failed to convince Z? If so, I'd be eager to learn what others do, short of using other programs, since I'm in too deep now to change. (I also posted a query to the Scrivener forum.)
  • Yes, Zotero devs have said this repeatedly and clearly.

    I don't know what other users do - I doubt the Scrivener-Zotero combination is all that frequent, not least because they have never integrated particularly well.
    If RTF-scan worked sufficiently well for you before, you should probably focus on getting that working again, starting with a more detailed report of what isn't working.
  • edited May 21, 2011
    How do you know this? Have others tried and failed to convince Z?
    It comes down to basic math. Zotero is a cross-platform app that runs on Windows, MacOS, and Linux, and has hundreds of thousands of users (not sure the exact number, but it's big).

    And yet, it's a project with limited resources: a handful of paid developers, and (unfortunately) not a big outside developer community.

    So it's not sensible for them to spend those limited resources on platform-specific applications with very small (relatively) user bases, when they are already supporting both Word and Open/LibreOffice.

    If I were you, I'd use the RTF scan feature.
  • This makes perfect sense. Alas, I haven't had good luck with the rtf scan feature - I followed all the steps, got the query from Zotero about each source I was using, but then I ended up with a final product that contained none of the cites, not even the curly brackets. It was as if I'd not entered cites at all! According to the discussion forum at Scrivener and a few here at Zotero as well, the rtf scan isn't working very well, even in the latest version.
  • I'm not quite sure I follow - you're "too deep to change" with two pieces of software you've never actually tried together before?

    Anyway, that Zotero would not write any field content is odd - did you try different citation styles? E.g. try with an in-text style like APA or APSA?
  • Over the years I have become increasingly disenchanted with Word, so I tried a number of things, including OO. But Scrivener won me over for the sheer pleasure of using it to write and organize as I was working on projects that didn't require citations. Meanwhile, I settled on Zotero about a year ago, and have been merrily gathering documents, taking notes, etc. for a new research project that does require citations. So now, for the first time, I want my two worlds to come together. It's like having two friends that I like and appreciate for their own qualities, only to discover that I had a nightmare dinner party upon introducing them. My post to the forum reflects my wish that I could maybe find a food they both like, ultimately with better results the next time they meet.

    I used Zotero's default, figuring that would be the safest.

    I think I might just write in Scrivener, using a few cryptic cites, export the document to Word, which I'll have to do anyway for publishers, then add the cites using the plug-in. It's all still faster than typing everything in!
  • edited May 21, 2011
    @cjkudlick
    Your initial request for more integration between Zotero and Scrivener: the good news is that it does work, although not perfectly.
    You might want to have a look at this and some of the linked posts over at the Scrivener forum. Although I share your particular cause you'll understand that it makes sense for Zotero and its volunteer team to improve the existing rtf-scan feature, which can be used in many apps beyond Scrivener, rather than come up with a plugin for a specific app.

    edit: fixed the link, thanks adamsmith.
  • kithairon - you forgot to include the link...
  • Link added.
  • edited May 21, 2011
    I looked at that thread. Just want to add that I think the OP's impulse is the right one: author in Scrivener, and use Zotero + RTF scan (once you/we get it working correctly!), and export to a form that Word can read (RTF, say).

    The problem with using Endnote or Bookends (what some on that forum are suggesting) is this will probably limit you going forward.

    I do think Zotero (both its data representation, and the rtf scan functionality) need to be improved to properly support this workflow. It's way too easy to have collisions just inserting an author and date, and so there really needs to be a way to store stable, human-readable, IDs, and to use those. There also needs to be support for th full range of features in the Zotero word/openoffice integration.

    Towards that end, I would love to see Zotero simply support the syntax developed by the pandoc community.
  • While this leads astray a little from the original issue - I mostly agree with bdarcus - I think adopting an already existing syntax is exactly the way to go.
    However, I would like to keep the option of using regular author-date format alive. I'm thinking of two different usage scenarios:
    Using labels is clearly superior if I'm authoring alone.
    But if I'm co-authoring with someone who isn't using Zotero or something directly compatible I'd still like to be able to tell him/her "just use author-date in curly brackets. with the following rules..."
  • Yes, so long as you have a way to disambiguate beyond author and year (say two or three words from the title?).
  • I haven't actually starting working on improving RTF scan yet, but I do still plan on taking this on. I personally don't like working with the Word/OO plugin, and would like to switch to the RTF scan workflow anyway.
  • @cjkudlick I'd have a simple solution for you, it is mentionned in the post on the website mentionned by kithairon

    Use a default output style in Zotero that generate autor-date with curly bracket like this one: {Autor, date} then manually add the page.
    http://rapidshare.com/files/449644793/RTF-Scan-Dragdrop.csl
    Then RTF scan and have your result display with the needed style.

    I wrote it down here basicly to have the info available on our forum so it can be found by search funtions.
  • MG6
    edited October 6, 2011
    For what it's worth, I once approached the Scrivener developer with this exact question, and he had much the same take on the situation (not encouraging!), but obviously from the other direction. He seems to be a fiction writer, and that is likely to remain his focus. He's nice enough, though. But given its flexibility as a text editor, I can see why many people would want to use it in combination with Zotero. Each is revolutionary at what it does.

    The problem I see, frankly, is that Scrivener's footnote implementation is awkward and pretty weak overall. And while RTF scan is adequate for simple citations, as I understand it, it can't handle multiple items, or prepended/appended text--is that so? If not, could someone please provide a link here to some relevant instruction? Either ability would go a long way towards solving my own issues with the combination of the two.

    Thank you!
  • you're right - RTF Scan is quite limited. I really hope this is something that will be taken on in the near-er future, but as the options for writing with Zotero outside of Word/Ooo are increasing - there is Pandoc with citeproc-hs https://github.com/singingwolfboy/citeproc-hs, there is zotero-plain: https://bitbucket.org/egh/zotero-plain/overview
    On the downside that means that a lot of possible outside developers have already found their preferred method of writing with Zotero and aren't likely to take this up. On the upside, it means that a lot of groundwork for improving RTF-scan is already done.
    All that said, I did think that multiple citations worked in the current rtf implementation. At least the documentation seems to suggest it does: http://www.zotero.org/support/rtf_scan
  • Ah... so it does, subtly. I didn't catch that. I'll play around with it. Thanks again.
  • FYI: because of the poor Zotero integration (in addition to a few other drawbacks) I switched back from Scrivener to writingoutliner.com, a very Scrivener-like plugin for Word that works well with Zotero.
  • And I bought an EndNote license, for a wonderful workflow on Mac:
    1. Reading with Preview
    2. Taking notes with DevonThink (I wrote a script that automatically insert the formatted and unformatted selected reference in EndNote (see here)
    3. Making draft with Scrivener (copying unformatted references from DevonThink notes!!!)
    4. Scanning document in Microsoft Word or OpenOffice

    I think this is the best workflow for academic writing!
    I'd use Zotero again if I have the possibility of applescript (for the step 2) and a good rtf-scan (for the step 4).
  • I don't know precisely what your script in step 2 is doing, but you can almost certainly replicate it by using Python (wrapped in AppleScript) and the libzotero Python library provided by (http://www.cogsci.nl/software/qnotero) for accessing your Zotero database. AppleScript access to Zotero would be an interesting project, probably best accomplished by writing a small shim program that used libzotero... Maybe I'll try if I switch back to Macs.
  • The step 2 is essential, because:
    - I can have the right template for different kind of reading notes (citation, paraphrase, summary, ideas). I can fill this template with a screenshot from pdf, or manual adding of text.
    - At the end of the template I automatically have the selected EndNote reference: formatted (Author, Title, Publisher, City, pages etc.) and unformatted {Autor, Year #ID @Pages}. The unformatted reference is very important because when I will work with Scrivener, I copy it in the footnote for example. I DON'T NEED TO SEARCH AGAIN AND AGAIN AFTER THE REFERENCE, LIKE IN ZOTERO!!! This is very nice and less time consuming!!!

    In the step 4, the scanning process will transform all these unformatted references in formatted references, according with the chosen style (APA, Chicago Style, etc.).

    I think this is the quickest workflow ever! Of course, you have to pay for EndNote, DevonThink and Scrivener. But if you are a researcher, it really deserves money.
  • I would strongly dispute the idea that EndNote deserves any money from anyone. It's an overpriced, mediocre product, behaves like a corporate bully (http://dltj.org/article/endnote-zotero-lawsuit-2/), imposes ridiculous EULAs on its users (same blog post) and is, after 25 years or so, still not able to export fully bibliographic data correctly in any open standard (see the various threads on importing from Endnote here).

    Go ahead and use it if that's what you want, but using extremely non-free products like Endnote imposes costs on your profession/community as well as on yourself by imprisoning your data in proprietary formats.

    (also, we're not deaf, no reason to yell).
  • edited November 27, 2011
    I agree with you. I love Zotero (I translated it in Romanian and I created two csl styles), but it can't replace EndNote in my workflow for the two reasons:
    1. rtf scan is feckless for serious projects.
    2. support for applescript doesn't exist
    For the first point, I've been waiting for more than 1 year. No hope. For the second I'm already hopeless…
  • edited November 30, 2011
    I've been a happy Zotero user for years, and I'm so impressed with Scrivener that I'm thinking of making it my main drafting engine. But getting Scrivener to play nice with Zotero is non-negotiable.

    I'm joining the conversation late, but everyone seems to be assuming Scrivener has to output an RTF file for Zotero to scan. But what about using Scrivener's MultiMarkdown feature with BibLaTeX?

    Zotero already has BibTeX citation export and can export references into a .bib database. Anjo7539 has written a Zotero to BibLaTeX translator. (MMD has its own citation format, but it's very primitive, so I won't discuss it further.)

    It seems to me that we're not that far away. We just need:

    1. A way to export BibLaTeX citations, with <prenote> and <postnote> fields

    2. A way to export the .bib file corresponding to these citations. Jason Friedman describes a way to automate this. (Will this work with Zotero Standalone?)

    3. If one is writing a specific document, one would simply add all its references to a collection dedicated to the document. This would then be used to create the .bib file.

    4. Ideally, there would be two enhancements. (1) Citation exports would be intercepted by a dialog allowing prenote, postnote, and author omission before pasting the citation into Scrivener; this is the plugin everyone has been asking for. (2) Instead of exporting a separate .bib file, a standalone program would access the Zotero library directly through Zotero's API.

    5. Another enhancement would be giving the user some control over the key format. Personally I've used the same algorithm for over thirty years and have had at most 2-3 problem keys. (first 3 letters of primary author's last name, first letter of last name, first letters of other authors' last names, 2-digit date, first major word in title; everything capitalized as they would be in a reference.) Zotero's key algorithm may be more foolproof, but it makes for awkward text.

    Most of this seems rather easy. I realize not everyone will want to use LaTeX. But it's not too hard to go from MMD to WYSIWYG formats, like Word or OOo.

    Am I missing something, or are we this close?
  • Just giving this a bump and subscribing, since I've been wanting to make the jump to Scrivener, but am waiting for better Zotero integration first.

    I'm hoping that what marsh says is true, and that it might be closer than I thought.

    I have some programming experience, although none with LaTex, but I would be willing to help out if there's anything I could do to help make this work. Thanks.
  • bump.

    #anti-segregation
  • There is really no reason to bump this - core devs are definitely not going to work on a specific Scrivener solution. They _might_, some day in the future, improve RTF scanning, but this is much more likely to happen if someone takes this on - as opposed to Scrivener, Zotero is open-source, so anyone with programming knowledge (javascript in this case) can take this on. RTF scan is a particularly good candidate, as it mostly happens in one single .js file. As I note above, some of the work has already been done as part of Zotero-plain.

    The same is true for the suggestion outlined by marsh above - anyone should feel free to start this.

    There is a fairly active and very responsive listserv for development issues, with questions you can always go there
    http://groups.google.com/group/zotero-dev

    (and to subscribe to this thread, there is a button on the left under "Notification".
  • I have not used scrivener, but if it can use extensions, it should not be difficult to port the Zotero word processor plugins to that platform.

    One way for a non-programmer to increase the chance that someone will pick this up is to make it as easy as possible. For example creating a wiki page that describes how you would like such integration to work (possibly with screenshots and sketches). Also you can point direct links to the Scrivener developer documentation that describes how plugins or extensions can be developed for that software.
  • Mronkko makes an excellent point.

    Is there a place documenting a Scrivener api or how to write extensions for Scrivener?
Sign In or Register to comment.