Ulysses III and Zotero - experiences?

I'm getting ready to write my dissertation using Ulysses III, a new, relatively unknown text processor that I used before to write another book. There is no Zotero plugin for it, but luckily with the RTF Scan feature, Zotero can be used with it, which is a boon. But I'm wondering: are others using Ulysses III (not the vastly different version 2) in combination with Zotero? I'd like to exchange best practice ideas, and if enough users would need a plugin or suchlike, we might petition either the Ulysess III developers or the fine people coding Zotero.
  • You're the first person to ever mention any Ulysses version here, so I'd say unlikely (though you never know), but do have a look at
    http://zotero-odf-scan.github.io/zotero-odf-scan/
    Even if Ulysses can't export ODF (which I don't know), you can export to RTF, open in LibreOffice, save as ODF and run the Scan. For a dissertation-length manuscript, I consider RTF-scan too inflexible and, perhaps more importantly, too unreliable, so I'd council against relying on that.
  • Please consider writing your doctoral thesis using LibreOffice. When writing and citing is complete and final (and you have removed the field codes) only then use Ulysses to refine and dress-up your manuscript.
  • edited October 20, 2013
    why? People find all types of different solutions that work for them. fbennett wrote his book in rsText, Rintze and I are currently writing in pandoc flavored markdown, we wrote the ODF scan plugin so Zotero would work better with Scrivener and Google Docs - the writing part is what's hard about a dissertation, people should use whatever works for them.
  • Thanks, Adam and DWL. I will look into LibreOffice and the ODF scan, but just to clarify: Ulysses is for writing many many words of semantically structured text, not for polishing the looks of a manuscript. That's for later.

    Please explain why the RTF scan would be unreliable. Does it simply not scale well?

    Thanks again for your comments.
  • Yeah, so my recommendation would be to use the cite-markers (what we call scannable cites) for ODF scan in your Ulysses document and then when you're done (and intermittently for testing) convert the document to ODF, scan it and create the final, lay-outed document in LibreOffice. I understand the appeal of lightweight, structured text editors for writing, so I wouldn't discourage you from that.

    Problems with RTF Scan:
    Won't recognize all citations (e.g. currently diacritics and umlauts in author names are a problem), specifying suffix and prefix for citations as in "(cf. Smith 1776, for an overview)" is impossible, it's hard (or impossible?) to disambiguate citations by the same author and year, it's impossible to specify different locators than pages (Smith 1776, chap. 3)
    One may be able to live with those limitations on 30 pages, but once you're looking at 200-300pages, it'll be a nightmare.
  • I'm simply recommending that, when working on something as important as a doctoral thesis, that tried and true software be used.

    I have used Ulysses (the current and an earlier version). I like it better than Scrivener although Scrivener is clearly more powerful. However, for a critical document that is likely to undergo multiple major revisions including moving large blocks of text, the risk of scrambling citations and bibliography entries are higher with these. I speak from experience as an adviser of grad students who have attempted writing their theses with these advanced tools. The prevalence of problems is more than 50% with Scrivener while with conventional word processors is no more than 10-15%.

    I make no argument that one should have the freedom to choose the software that works for them. I still say that for what is perhaps the most important document of one's life, it is prudent to use something that has been well-tested and with which the writer has had much experience.

    Sebastian, I greatly respect your mind as a coder. I don't know the software you used to write your dissertation. However, you and Frank Bennett, are not typical writers. I believe that the typical writer of a thesis is more likely to need to think more about the topic and citing authority than the process of writing.

    Writing with Ulysses III is a splendid experience. I have memories of what was good about WordPerfect when it was DOS and worked with Reference Manager. However, it is to me what it claims to be -- a tool for creative writers. It is not designed for academic writers. The multiple step process required -- exporting in (questionable) RTF, importing and converting to ODF, and then going back to Ulysses -- provides opportunities for errors and nightmares.
  • I still say that for what is perhaps the most important document of one's life, it is prudent to use something that has been well-tested and with which the writer has had much experience
    Yes, and given that rhvdg says they've written a book using Ulysses, let's assume they have good reasons for wanting to continue using it and stick to the question at hand, which is about using it with Zotero.
    The multiple step process required -- exporting in (questionable) RTF, importing and converting to ODF, and then going back to Ulysses -- provides opportunities for errors and nightmares.
    No one is suggesting going back and forth between programs. The whole point of RTF/ODF scan is that you can write in whatever text editor you prefer and generate a final formatted version at the end. That may not be the workflow for you, but there's no need to derail discussions among others about doing so. (And unless you have experience generating malformed RTF from Ulysses, the part about it producing "questionable" RTF is just FUD. With plain-text source documents, automated backups, and test conversions along the way, there's not a whole lot of risk here.)
  • Thank you, all, you've been very helpful. I see the value of the ODF Scan approach, even if until now my testing of the RTF Scan didn't reveal major weaknesses. My field is philosophy and my dissertation will concentrate on a relatively small number of sources, which perhaps makes the ODF route less of a necessity. But being able to add text like "See also" and "chapters 3 and 4" to a citation is a boon, to be sure.

    The Ulysses III developers seem to be considering Sente support. I hope to convince them to look into Zotero first. Ideally, UIII would "package" the ODF Scan markers, like URLs and footnotes are now "packaged", in nice little softly coloured buttons hiding ugly markup while showing what's essential, in this case author and year, and .odt would be added to the list of supported export formats.

    (In the very best of all possible ideal worlds, inserting citations would be a matter of a keyboard shortcut causing a floating window populated by available sources to appear. Type the first few letters, click, done. Many use UIII full-screen, so dragging citations in or even selecting them in a different application would mean stepping out of the distraction-free environment... Dangerous! Anyway, one can dream.)
  • ODF Scan is still an unofficial extension, and there might be changes we would make before integrating it directly into Zotero as a replacement for RTF Scan, so it's probably premature to try to add support for it to Ulysses. But given that personally I couldn't imagine writing a document of any significant length not in a plain-text editor, I'm interested in our having better official support for this workflow eventually.
  • I am in a similar position. I'm also a philosopher. rhvdg, let me know if you've made much progress on this issue.
  • I looked into using the program in the winter, but found it was both more and less than I need. I've ended up writing everything using TextMate in Pandoc's Markdown. (Pandoc, by the way, is written by a philosophy professor, and has a particularly nice way of processing citations.) I use zotxt for integration with Zotero. Works flawlessly (with the exception of a few known weak points in Zotero itself).
  • I've spent days trying to figure out a way to use Ulysses and Zotero together. Best I can come up with is... can't really.

    I basically write in MD using my own made-up cite keys in the text. Then when draft is done, I export to get manuscript into Word, and then change out citations one-by-one using the Word plugin for Zotero... then generate Reference list.

    Yuck.

    Which there was something like ZotQuery that was just a tad more user friendly for non-programmers. the Alfred app plus power pack is a lot of $$ to get into to just try out Zotquery. I can read that ZotQuery will produce an alphabetic list of works cited by basically spitting out a Zotero collection, but can it do the equivalent of the RTF scan and produce a reference list in order of cited and re-write cite key with numbers as per style required?
  • if you're writing in MD anyway, why not give pandoc/citeproc-hs a try? That seems a lot better then going the Word route.
  • I am hungry to try pandoc, but the command line only interface is intimidating to me. Waiting for a nice GUI with switches and input/output options...
  • So you've explored ODF scan plugin and it didn't work out for your Ulysses workflow?
  • edited June 15, 2014
    Honestly, the terminal isn't as scary as you might think, and you certainly don't need to be a programmer (see this tutorial and this for basic information). It allows you to accomplish a number of things much more efficiently, and opens up an entire world of small, useful utilities. Many of these (including pandoc) can be installed very easily on the Mac using Homebrew. (For instance, to install pandoc and its citation processor, you simply type 'brew install pandoc-citeproc', and it takes care of the rest.)

    That said, yes, it would be nice to have a Markdown editor that also acted as a graphical interface for Pandoc (and, of course, included a version of the Zotero word processor plugins that would insert a citation key). Still, TextMate or even TextWrangler allow you to do much more than you can with a typical word processor.

    The wonderful thing about the combination of pandoc and Zotero is that it's the first usable option I've found for doing writing that is truly portable, and allows me to focus on the writing rather than its formatting, which (as I have learned the hard way) publishers just don't generally care about, but they often do have specific requests for providing your text in a specific format (usually Word), and pandoc makes it extremely easy to provide whatever is needed. It's also been really useful for teaching, since I can make up a handout (e.g. collection of a primary source texts) and give students a PDF and EPUB version (which students have really appreciated) from a single source, plus make up slides from it (using pandoc's export to reveal.js format) by simply adding a few horizontal lines.

    The other great thing about working in plain text is that it allows you to use a version control system, which (despite their typical limitations for writers) I find is both a very nice way of keeping track of what you're doing with a specific writing project and a great way of managing revisions without having to keep multitudes of old files.

    There's all sorts of other information about this online; this post by W. Caleb McDaniel is a good introduction.
  • Thanks, Dunning. Despite reading everything I could find, I didn't know it could be that easy to install pandoc. Will give it a try. I agree with the rest, and thus why I'm trying to forge a reasonable workflow using Markdown. Roundtripping everything through a Word install at the work office has made me crazy.

    What are people using it add Zotero cite keys then in a Pandoc format? Zotquery?
    And to format a reference list?
  • What do you mean by reference list? pandoc/citeproc-hs will automatically generate reference lists for items cited in the text, so no need to have a special way of doing that.

    If you just want a reference list you might as well do it in Zotero, and create it as RTF or html, no? If you_really_ want it in MD, you can generate it in html and have pandoc convert it to MD, but I don't really see why that would be of interest to anyone.

    @dunning - I'd be curious what you use for the citekeys, too.
  • edited June 15, 2014
    I use Erik Hetzner's zotxt: it adds an 'Easy Citekey' translator to Zotero that will output a citation key in the format of @doe:2014title (you can change the key to something else by adding a note or tag to the item starting with @). You then simply run pandoc with the accompanying script, 'pandoc-zotxt', which automatically passes the bibliographic information from Zotero to pandoc-citeproc for the items cited. It works extremely well.

    (The same author has also written zotxt-emacs for autocompletion of citekeys, which might be of interest to some here, but, not being an emacs person, I haven't used it myself.)

    See the page on pandoc's citation format for more details on citing items and getting a reference list. In case you use footnotes, one thing it doesn't make totally clear is that it essentially expects you to cite everything in an author-date style; it automatically moves the punctuation as necessary when using a note-based style (so a citation placed before a period will result in an inline reference in brackets before the period, but a footnote reference after a period). This allows you, for instance, to output an HTML version of an article using Chicago author-date style plus a PDF using Chicago full note, both using the same Markdown file. It's absolutely brilliant.
  • this is all brilliant.

    Thanks @adamsmith. I guess I will read more on how to generate the reference list using pandoc. You're right, no need for it in the MD stage.

    @dunning, came across zotxt today as well and intimidated with install and usage. Will push to try it out.
  • You simply install the zotxt add-on from the Mozilla website into Zotero and paste 'sudo pip install pandoc-zotxt' into the command line. That's it; there's nothing to configure (except that you need to add the extra '-F pandoc-zotxt' bit when processing a file with citations from Zotero, as described on the website).

    I hear where you're coming from, as I worked with an assumption for years that if something required the command line, it probably wasn't worth using. But many of the best programs out there are based on these sorts of open-source workhorses that focus on the quality of their output rather than fancy interfaces. (Compare Word for Mac's almost complete inability to do something as basic as produced a reasonably-sized PDF.) Pandoc is so good that I can't imagine there won't eventually be a graphical wrapper for it, but in the meantime it's an extremely efficient tool for academic writing (and for publishing in general).

    I should also note that a docx reader was just announced on the pandoc-discuss list, meaning that pandoc will have built-in support for converting Word to Markdown (it's possible right now using other tools, but it's a pain). This will only make things more interesting.
  • excellent coverage of this... thanks.

    I feel like I've come a long way, but now facing a mile of Python errors on first run of Pandoc. May need to bench this and circle back in a few days.
  • That's odd. Are you running the current version of OS X? Not sure if that would make a difference. Note also that you need to have Zotero open in order for zotxt to be able to import the references from it.
  • edited August 16, 2014
    I'm a bit late to the discussion, but I would like to share how I use Zotero and Markdown/Pandoc in Windows (it is useful for Mac users too I believe). The zotxt approach sounds great, but using pip to install packages is just a pain in Windows (at least I gave up on it).

    My solution is to use Better Bibtex (https://github.com/ZotPlus/zotero-better-bibtex) and Zotero autoexporting (https://addons.mozilla.org/en-US/firefox/addon/zotero-bib-autoexport/).

    After installing Better Bibtex, go to its preferences and set the following citation-key format: [auth][year][Title:skipwords:select,1,1:capitalize:nopunct]

    This imitates the citation key system of zotxt using authoryeartitel (everything uppercase), which I find easy enough to remember and still unique enough to identify the items I cite. With Zotero autoexport I regularly update a Bibtex file containing my whole library.

    In my writing, I use Pandoc's citation syntax together with my citationkey. When I'm done writing and ready to export, this is my workflow:

    - open the folder with your writings (a good folder structure is king when you write with markdown)
    - hold the shift-key and right click in the Explorer. This gives you the "Open command window here" option in the right-click menu
    - use pandoc with the --bibliography command, like this (here your bib-file is in your documents folder):

    pandoc markdownfile.md -o markdownfile.docx --bibliography C:\Users\<Username>\Documents\Library.bib

    Also, when you use Windows, there is only one editor that I found really suitable for writing: Sublime Text. Just forget about the rest. Here is a great manual for how to set up Sublime Text for Markdown editing: http://plaintext-productivity.net/2-04-how-to-set-up-sublime-text-for-markdown-editing.html

    Edit: I updated the citation key format for Better Bibtex. And good news: The newest version of Better Bibtex updated drag and drop citations for Pandoc, just select 'Pandoc Ciation' in Preferences -> Cite
  • Thanks, @S.Baack, for your comments; I was able to modify your citation key format to match zotxt exactly:

    [auth:lower:select,-1,1]:[year:nopunct:select,1,1][Title:lower:nopunct:skipwords:select,1,1]
  • You're welcome, @dunning, glad you found it useful and thanks for sharing your modification. Personally, I don't like the colon in the zotxt citation key format, which is why I didn't include it.

    Meanwhile, I found another great Sublime Text plugin that makes my system more comfortable to work with: Citer (https://github.com/mangecoeur/Citer). It provides citation search and Tab-completion for citations stored in a bibtex file. Inserting citations in plain text Markdown is now just as comfortable as in Libre- or MS-Office (with the Zotero plugins).
  • edited October 31, 2021
    Ulysses works best for me to edit text, then I convert to HTML to be posted on a webpage. I'd love to have a button that (like what lives in the Word-Zotero plugin) that lives in Ulysses--where I can just slam the reference in the document without jumping around and clicking buttons and saying incantations (what it feels like when I read the above descriptions). I also found Scrivener and Word to work best after you have a good first, second, or third draft. Considering the popularity of Ulysses...hopefully, my fantasy will come true.

    As an addendum...Zotero for me has been close to life-changing. I started a group of over 4,000 physicians (the Cellular Medicine Association)--I preach Zotero to anyone who will listen as the way to the Kingdom (from Maddening Frustration to FUN with references). For Books, I'm using Ulysses, then for the final format in Word, everything is perfect...because it all stays in Word when it goes to the publishing software. And, all is perfect when adding a bibliography to a website, I select everything I want and just slide it in one bulk slide from Zotero to the webpage (it's a miracle). But, when I use Word and then try export as HTML, all the lovely numbered footnotes lose their formatting...so I'm left with only a non-indexed bibliography at the end of the page. Maybe this addendum just confuses the issue...bottom line is that if I had the same functionality in Ulysses that I have in Word, that would be best for someone who glazes over when using Zapier or Alfred.
  • Word integrates with Zotero with the elegant Zotero button that appears within the Word software: click the Zotero button while working in Word, choose the reference, click, and the footnote is added and the bibliography updated. When will there be something as elegant for Ulysses?
Sign In or Register to comment.