Papers 2: the reference manager made with love

The reference management software Papers has been a regular topic on this blog. I wrote about Papers in one of my first blog posts in May 2008, interviewed main developer Alex Griekspoor in October 2008, and held a poetry contest when Papers for iPhone was released in February 2009 (Stephen Curry won the first price for this poem).

Today Papers 2 was released, and with this major update Papers has grown into a full-fledged reference manager for the Macintosh (there is no Windows or Linux version). Papers 1 was released in 2007 as a tool to manage the PDF files of scholarly papers on your hard drive. Papers 1 did this job very well, but it was not a reference manager in the strict sense, because you couldn’t use Papers to insert citations into the manuscripts you were writing.

In order to add this feature, Papers first had to be completely rewritten to not only handle journal articles and the associated PDF files, but also all the other common reference formats – conference proceeding, book chapter, website, etc. A nice side effect that I haven’t fully explored yet is that Papers can now also store all your Powerpoint or Keynote presentations (but fulltext search in these presentations doesn’t seem to work yet).

Inserting references into manuscripts is done using a floating window activated from the menu bar or with a shortcut:

This feature works similar to the plugins used by Endnote, Mendeley, Zotero, etc., but in typical Papers fashion the implementation is much nicer. The plugin is available to all applications, e.g. other word processors besides Microsoft Word, your blogging software or your email program.

The other big limitation of Papers 1 wasn’t so obvious when Papers originally launched in 2007. But today most major reference managers allow users to share their references in private and/or public groups, as most research is done in groups and most manuscripts are written by multiple authors.

The sharing feature of Papers 2 is called Livfe. It is not as complete as the sharing features of some other reference managers (e.g. CiteULike or Mendeley). It is for example not yet possible to upload a complete Papers library to Livfe. And Livfe doesn’t have a web interface, but only works through Papers.

The difference between Papers 2 and other reference managers is not so much the feature set. What sets Papers really apart is the love that went into the design of the program. Papers doesn’t just get the job done, it allows you to have fun searching for references or reading – and now also writing – a paper. This makes Papers 2 a strong competitor against some of the other established programs that for example have more than 25 years of experience with reference management (Endnote), think that reference management should be done as Open Source software (Zotero), do everything in the browser (Refworks), or see reference management primarily as a social application (Mendeley).

This entry was posted in Conferences, Interviews, Presentations, Recipes, ResearchBlogging, Reviews, Snippets, Thoughts and tagged , . Bookmark the permalink.

35 Responses to Papers 2: the reference manager made with love

  1. Mark says:

    Why cant/wont they make a windows version. Surely this would be in their financial interests?

  2. Martin Fenner says:

    Papers is very much a Macintosh application, and it takes advantage of many unique features of that platform. I don’t think that the Mekentosj folks care very much about the increased market that the Windows (or Linux) platform would offer.

  3. Jongo Gurmola says:

    They don’t care about octupling their effective market/income? That’s a foolish business decision.

  4. Gonzo says:

    Doubtful. First off, the Windows market is already more heavily populated with reference managers, so the competition would be greater. More importantly, though, Mac OS X has a much greater market share in academia than than it does in the business or general consumer population. In my institution (large American university pulling in something on the order of $1.5 billion/year in research grants) Macs probably outnumber Windows machines by a factor of 2:1; in my department it’s more like 10:1.

  5. Bruce D'Arcus says:

    Martin, on this:

    This feature works similar to the plugins used by Endnote, Mendeley, Zotero, etc., but in typical Papers fashion the implementation is much nicer.

    I find this is a really odd way to characterize the comparison. Papers2 seems to make entering references easier (though can you do local modifications like page locators, author suppression, “see” prefixes, etc.?) in a wider range of applications (though limited to a single platform), with one pretty major limitation: they’re not live citations! Creating finished documents requires creating a separate document.

    So rather than make a blanket statement about which is “better” on this front, I think it’s more accurate to say they each have different trade-offs.

    Also, worth noting they’re using CSL for formatting.

  6. Mike Fowler says:

    Thanks for the background, Martin, and for the other info from other commenters.

    I have Papers 1 installed, but I’ve never actually bothered to use it, as it was too limited in the CWYW options to be more useful than Zotero/Mendeley as a reference manager. I’m glad to hear Mekentosj have updated to incorporate this, but would appreciate some feedback on the breadth of output citation styles that they offer bundled with the software – a wide range of journal specific options like Endnote offer, or a limited set of generic options like Zotero and Mendeley offer?

    Related to that, if it’s a small, generic output list, is there a large enough user community to supplement this, as the freeware options can rely on?

    As for business models, it’s obviously been successful enough for them to invest in R&D for this upgrade.

  7. Pingback: Weekend shorts (including Japan earthquake links) | Code for Life

  8. Bruce D'Arcus says:


    I’m glad to hear Mekentosj have updated to incorporate this, but would appreciate some feedback on the breadth of output citation styles that they offer bundled with the software – a wide range of journal specific options like Endnote offer, or a limited set of generic options like Zotero and Mendeley offer?

    A few things:

    First, Zotero, Mendeley, Papers2, and a few others apps all use the same style language (CSL), and it bothers me a lot that people don’t know this. It doesn’t help that commercial players in this space seem not to be very good at advertising this fact, merely trumpeting they can all read the same 1000+ styles, without explaining how. Maybe I need to find some designer to create a nice little CSL logo so people can use that and instantly indicate their apps are CSL compliant :-)

    Second, you’re also a bit off-base on the assertion there’s a “small, generic, list of styles.” As I indicated, there are more than 1000 (last I checked, it was more like 1500).

    Related to that, if it’s a small, generic output list, is there a large enough user community to supplement this, as the freeware options can rely on?

    Now this is an interesting question. Here’s my hope:

    We need a really nice, beautiful, online repository and style creation and editing app that users from all of these apps can benefit from and contribute to. So if a Papers2 user needs a new style, they should be able to go to this site, and in two minutes have a style they can use by merely clicking a link. That style should instantly be available for a Zotero or Mendeley user.

    The question, is, though: how do we make this happen?

    One possible answer: commercial players that are benefiting from this ecosystem should contribute resource (programmer time, say) where they can. Mendeley has already stepped up; would love to see the others.

  9. Mike Fowler says:


    As I mentioned in my 1st post, I was asking about the styles bundled with the software. Zotero (my current reference software of choice) has a more complete list of CSL styles, mostly contributed by users, available here, though they’re not all perfect, and it’s far from complete for, e.g., ecological journals (my field).

    I’ve tried modifying some existing styles with a web based CSL editor (sorry – don’t have the link at the moment, but it can probably be found through the Zotero pages), but it’s usually a kind of temporary hatchet job, and I don’t feel confident submitting my dodgy styles to a wider user community.

    I like your suggestions, Bruce. That would really benefit a lot of people in a lot of ways!

  10. Martin Fenner says:

    Mike, as Bruce has said, Papers 2 uses the Citation Style Language, specifically CSL 1.0. This CSL version is supported in Zotero starting with 2.1, the CSL 0.8.1 used in earlier Zotero versions is not compatible (Bruce could explain the details). Papers 2 unfortunately doesn’t include a CSL style editor. I asked the developers, and there are unfortunately no plans to add a CSL editor in Papers 2.1.

    CSL is clearly the future for citation styles. is the best starting point and I hope that in the course of 2011 we have both a growing number of CSL 1.0 styles, and a CSL editor that people can use to contribute additional styles.

  11. Martin Fenner says:

    Bruce, I will rephrase my comment about Papers 2 and how it inserts citations. The user interface of the implementation is very elegant and is fun to use. This doesn’t say anything about the supported features. As I said at the end of the post, Papers 2 is a good choice for users that care about usability. The competitors of course have other advantages.

  12. Bruce D'Arcus says:

    I’m not trying to criticize Paper 2; more trying to emphasize that this stuff is hard, and that each product tends to prioritize different functionality and workflow. So I would really like a quicker way to enter citations in Zotero, for example. But I really can’t do without the advanced citation features it has (note: I don’t know that Papers 2 doesn’t also have these; just a hunch that it doesn’t based on what I’ve seen), nor would I be willing to do without “live” citations (for that workflow, I use markdown and emacs ;-)). To me those are about “usability” as well.

    But I take your point.

    On the style editor, I really think the emphasis should be on the finishing the web-based app that Mendeley has sponsored. It’s really hard to do this right, as CSL is more complex (to accommodate more features) than earlier examples like Endnote’s styling language. So it’d be better if different CSL implementers dedicated resources to finish what is a promising start.

  13. Matthias Wucherer says:

    @Jongo Gurmola, “They don’t care about octupling their effective market/income? That’s a foolish business decision.”:

    Sometimes it’s not always about the perfect economic decision. Sometimes great products arise from minds having great ideas, making something with full enthusiasm, not so much out of direct benefit. Did you ever wonder why it is small garage enterprises that have the great ideas and inventions, and big companies simply buy these ideas because they don’t have such ideas themselves because for economical reasons they cannot make these “foolish business decisions” in the first place?
    If all developers would think this way, an application like Papers would never have been developed in this streamlined, strictly science-focused, non-compromised way.

  14. Sean Takats says:

    The reluctance to acknowledge CSL more openly has undoubtedly limited the quality and quantity of available styles. Zotero has the largest and longest-standing user base of any CSL-based software, but it still bears a disproportionately large share of the development of styles, which is to say, all of it. In fact, I’m not aware of a single style that has been contributed or revised by a user of software other than Zotero.

    That the hypertrophic user stats trumpeted by Mendeley have resulted in no community involvement is troubling, to say the least. I would suspect that some of the disparity is attributable to the maturity of the software and its community: while researchers have had years to write dissertations, journal articles, and books with Zotero, they may only be just beginning to do so with Mendeley. Nonetheless, there are deeper problems. For example, the support and maintenance of CSLs has largely been outsourced to the Zotero community, through the Zotero forums. The WYSIWYG style editor project, while laudable, is alas moribund. And without a concerted effort to engage a user community in the creation and refinement of styles, the editor in and of itself would not seem to provide any real contribution toward improving the number and fidelity of CSLs.

    That Mekentosj developed its CSL-based citation engine in complete secrecy is again, weird and maybe a little troubling. One accusation regularly leveled against Zotero is that it’s “free and open-source,” as if these were somehow dirty words impugning the software’s quality and reliability. But to the extent that an entire ecosystem of tools now relies upon styles that are community-built — and in the case of Mendeley, the formatting engine and word processing plugin code, too — maybe it’s embarrassing to admit that the most essential guts of these projects, the parts that really make them work for real researchers like me, are not self-built.

    Unfortunately, I don’t see how the projected CSL style repository from to will do anything to solve this problem, but we’ll do it, because, why not?

  15. Rintze says:

    @Sean Takats:
    Mendeley is not really a new kid on the block anymore. But Zotero’s continuing dominance in CSL development is perhaps not so surprising: Zotero was the first reference manager to adopt CSL, and Zotero’s ecosystem is simply more inviting to outside contributions: Zotero users can, and do, contribute web translators, client code, styles (through the repository), wiki documentation, plugins and translations. All help in creating a robust user community. And the Zotero forums and mailing lists are better suited for dialog than e.g. Mendeley’s feedback system.

    I’ve been pondering if it would be a good thing for the CSL project to set up its own forums. But while it might make sense in the long run, I’m a bit afraid we’ll lose part of Zotero community in the short term.

    As for Mendeley’s CSL WYSIWYG editor, you’ll probably recognize that it’s not an easy project. Big challenges are visualizing CSL’s hierarchical XML in a clear way in the GUI, and helping style editors with finding a good starting point for style development (almost all existing CSL styles are derived from other styles).

  16. Martin Fenner says:

    What is the next step for CSL then? Do we first need a working web-based citation style editor? And how can one contribute to the development? And should we ask developers of commercial reference management software using CSL to contribute more to this?

  17. Rintze says:

    While a CSL style editor is certainly desirable, I’d say the first priority is setting up a CSL 1.0 style repository (with previews and an easy way to browse styles) and figuring out the details of distributing and contributing styles. We made a start by creating a github repository ( for CSL 1.0 styles, but especially for building the front end, contributions from developers of (commercial) reference managers are more than welcome.

  18. Mr. Gunn says:

    Mendeley users are just beginning to write styles. I field tons of complaints along the lines of “Style X is wrong! How could you possibly expect to be taken seriously by anyone in the field of marmoset sociology when you can’t even format citations properly?” I always explain about CSL, that’s it’s their colleagues and other scholars who are creating these styles, and then point them to I also often hear grumbling that if a journal wants their own style, the publisher should write the CSL file and make it available, instead of expecting authors to do it. Mostly the style effort from Mendeley users has taken the form of very limited fixes/updates to existing styles, and I think a CSL editor that’s easy to use would go a long way to getting involvement from a wider group of people. However, I’ve suggested to some of the people who I know have fixed existing styles that they submit their version back to the repository, and they have all found the existing process fairly cumbersome.

    Given all this, I would agree with Rintze that the first step would be figuring out the details for contributing and updating styles to a shared repository. I don’t think we’ll see a level of contribution of styles commensurate with the size of the Mendeley userbase until the CSL editor is widely available, but when we do, we want to make sure all that effort is captured and made available for everyone in the CSL using community.

  19. Luke Dones says:

    I have Papers 1.9.7 and would like to export my references (almost 400) to Papers 2.0.1. How does one do that?

  20. Martin Fenner says:

    Luke, wenn you use Papers 2 for the first time, it will ask you to import your Papers 1.9.7 library.

  21. Martin Fenner says:

    Rintze and William, I agree that an easy-to-use CSL editor would be helpful, but that a central CSL style repository is even more important. It would of course help if publishers would collaborate and contribute styles (and link to them in their author instructions). If a journal, conference – or professor at your university – wants you to use a particular citation style, they should provide that style in CSL 1.0 format.

    Now that Zotero 2.1 with CSL 1.0 support was released, I wonder when the Zotero Style Repository will list CSL 1.0 styles instead of CSL 0.8.1 styles, and how the Zotero and style repositories relate to each other.

  22. Luke Dones says:


    It didn’t do that, but I will try again. Thanks.

  23. adam.smith says:

    Mr Gunn:
    in case you’re still reading this – submitting a style for the csl repositories (currently both Zotero and the one set up via github) doesn’t take more than putting the style online and writing a message to the Zotero forum. We’ve recently clarified the instructions on this.
    Please point people to those instructions.
    I tend to agree with Sean – as of now, you guys are essentially free-riding on the Zotero community for this, but I certainly would be happy to see this change.

    The hope is to move the repository – though likely not the main forum for style development for the reasons that Rintze points out – away from Zotero to entirely.
    I don’t think the Zotero repository should even be updated.

    I do most of the style uploading over at Zotero and once there is a working repository at I’ll work with that exclusively – right now I’m working with both repositories (and two different submission systems…) and I don’t intend to do that much longer than a couple of months.

  24. Mr. Gunn says:

    Thanks, Adam, I’ll be sure to point people to that. Of course, we appreciate y’all taking the lead on CSL development and getting the community as far as you did as fast as you did. Expect much more from Mendeley users in the future.

  25. Martin Fenner says:

    Adam, thanks a lot for the info.

  26. Bruce says:

    And another update, as I missed much of this.

    As you all know, there is a github repo.

    I have also made clear that I’m going to be quite liberal about adding style editors to manage it, and that I encourage pull requests. As I’ve yet, however, I have zero offers to help with this from anyone but people from the Zotero community.

    It’s really time for the commercial players to start stepping up. Free labor only goes so far (in particular, those of us dedicating it easily get burned out, have other priorities, etc.).

    I’m optimistic about the editor, but a) as Rintze says, I’m not so sure it’ll work if we can’t even get basic repo coordination settled, and b) as I’ve repeatedly said, the devil is in the detail on the editor; if it’s not really good, it may lead to a lot of really bad styles.

    And yes, it’s about time the journals, which write these insane requirements, to start hosting their own styles!

  27. Pingback: Literaturverwaltung neu gemacht – Eine kleine Zusammenstellung aktueller Softwareentwicklungen (Teil 1) « Literaturverwaltung & Bibliotheken

  28. Remberto says:

    Short/simple question:

    -How can I edit my citations, e.g. add the page number
    This is basic stuff and I can’t manage to find it in Papers2.
    Only can I enter the general work, paper, book, etc… but cannot add the page from where I found the sentence/idea I wanted to cite.

    How can I add the page number to the citations?????

    THANK YOU!!!

  29. Remberto says:

    Short/simple question:

    -How can I edit my citations, e.g. add the page number
    This is basic stuff and I can’t manage to find it in Papers2.
    Only can I enter the general work, paper, book, etc… but cannot add the page from where I found the sentence/idea I wanted to cite.

    How can I add the page number to the citations?????

    THANK YOU!!!

  30. Martin Fenner says:

    Remberto, you can add page numbers and other information in the “Info” window on the right. This should rarely be necessary for journal papers, as their citation usually can be imported from online sources.

  31. Remberto says:

    Thanks, but I meant from Word processor apps. As a citation manager.

  32. Martin Fenner says:

    Remberto, do I understand you correctly that the page numbers are in Papers 2, but you don’t see them when you insert a citation into the word processing document? If this is the case, then you should try a different citation style. APA and Vancouver are two popular styles that show page numbers for journal articles.

  33. Remberto says:

    I know. But try inserting the page number with papers2….. good luck!

  34. Kaz says:

    I think what Romberto is trying to say is that he wants to cite the page number when inserting citations from Papers 2, like (Author 2002, p.299), which is something I’d like to do as well but can’t quite figure out how to do it.

  35. Karl says:

    Like Remberto and Kaz I would be very interested to know, if Papers 2 is able handle page numbers within citations. This is something I need to use nearly every day. I tried to insert it by myself (by manipulating the cite-key), but it didn´t work.