As I mentioned in a previous post, reading and writing are central to my academic career. I am a voracious reader. I not only have to keep current in my clinical specialty, Anesthesiology, but also in my research areas, Simulation and Games Based Learning. Because of this, I’ve developed multiple workflows that make reading and annotating more enjoyable, efficient, and useful.

I’ve mentioned my preference for Papers2 app to keep track of scientific publications. What I didn’t mention is that I’m a prodigious highlighter-it is not uncommon for me to mark up a single manuscript with dozens of highlights and notes. Keeping track of the information in manuscripts AND my thoughts about what I read used to be a real chore. Before Papers included highlighting, I had a complex workflow to accomplish this feat. Now it’s relatively simple.

As I mentioned, when working with scientific literature, I like to batch operations. I go on paper collecting binges where I do little (if any) reading. Only after I’ve completed my collecting, is it time to read. I actually schedule time on my calendar to read, annotate, and synthesize ideas.

This entry is about my annotation workflow–and how I use technology to speed up the process of retrieving concepts when writing and citing.

I do most of my reading (and annotating) on one of two devices: my Mac or my iPad. I read so much, it’s hard to keep track of all the ideas. There are few things more frustrating than knowing I’ve read something, but being unable to find the source. This workflow alleviates the stress of finding relevant information in my library. But even more than that, using artificial intelligence in a tool called Devonthink, I can find links between concepts that were not initially apparent to me.

This workflow has been a personal success. Before implementing this workflow, I could spend hours trying to track down an appropriate reference. Since implementing this workflow the same task takes seconds.

I also wanted to thank Derek Van Ittersum (Kent State | blog) for inspiring me to polish this workflow–I had a chance to collaborate with Derek on the Mac Power Users 100 show.

Programs needed:  

Setting Things Up

I’m assuming you’ve already installed DropBox on both your Mac and you iPad. As I discussed in a previous entry, I configure Papers save PDFs to a designated folder Dropbox. I set this up by navigating to the preference menu in Papers and designated my preferred folder you in Dropbox. I do this to have the PDFs in my Papers collection accessible from anywhere and by any tool.

How Papers handles PDFs

Within the Papers preferences, I’ve configured Papers to launch a PDF in a new tab (using Papers itself, NOT an external PDF reader).

Launch PDF within Papers


Papers on the Mac and iPad have built in highlighting tools (on the Mac, invoked by hitting the control key over highlighted text, or on the iPad by holding a finger down and dragging). As I highlight or make notes in Papers , a separate layer is created in the PDF that keeps track of annotation information. As I annotate, each highlight and note appears under the “Notes” tab in Papers.

Notes Tab in Papers

If I find a paper I’d like to read that’s not in my library I copy and paste the title and author to my OmniFocus Inbox-to be gathered during a future collecting binge. This way I don’t interrupt what I’m doing.

After I’m done highlighting and note taking, I write a single summary note that captures the essence of the paper. At this point I also tag the paper Papers Keywords. Once I’ve completed my summary, if I’ve been reading on my iPad, I synchronize the PDF back to Papers on my Mac. After syncing, the annotations I made on the iPad are available on the Mac (and visa versa). If I’ve been reading on my Mac, I don’t need to synchronize to perform the next step.

Once the annotated PDF is in the Papers collection on my Mac, I go to File:Export:Notes within Papers (make sure the “export selected paper” is highlighted and RTF is selected) to export a Notes file to a folder on my Desktop. RTF format is important–Devonthink relies on it.

Export Notes from Papers

A limitation of Papers is that it can’t export each note or highlight separately–it dumps all my annotations into a single file. I open the file and cut and paste each individual note (that represents a single highlight) into its own file. I then comment on the highlighted passage. Finally, I use the “Magic Manuscript” feature of Papers (invoked by hitting my Option Key twice) to append the citation reference to each statement.

Extracted Note with Magic Manuscript Citation

Finally, I drag each RTF into to the “Supplemental Files” tab of the Papers. This saves each comment in a supplemental folder residing side-by-side with the manuscript in the Papers Dropbox hierarchy.

Adding RTF to Supplemental Files in Papers

The next step of the workflow uses the “Indexing” feature of Devonthink Pro Office. Using Devonthink, it is possible to index (or reference) any folder on your hard drive. This makes the contents of PDFs and RTFs available to Devonthinks Artificial Intelligence without directly importing the information into a Devonthink Database.

I created a Devonthink Literature Database that indexes (does not import) my Papers2 Folder. I did this the first time by opening the File:Index… Menu item in Devonthink and navigating to the Papers Folder in Dropbox. All subsequent updates are done by opening my Literature Database, highlighting the indexed “Papers2” Folder, and navigating to File:Update Indexed Items. Indexing can takes quite some time–be patient.

Devonthink Indexing

This all sounds complicated, but trust me, it’s worth it. What this allows me to do is open Devonthink and find any statement, concept, or related item quickly. I can then cherry pick comments related to my search and paste them into a draft document with little modification.

When I’m writing, I take all related concepts and paste all of them into a single card in Scrivener. This allows me to arrange, and rearrange information to help support my argument. Because I took the time up front to include the Magic Manuscript Index from Papers, no further searching or citation work is needed, until I’m ready to format my bibliography.

Scrivener for iOS is under development–it is likely my workflow will change when it is released. Instead of using Scrivener, one might consider using Tinderbox –Derek’s tool of choice or the newly announced Scapple (by the makers of Scrivener). I will cover the next step of the writing process using these tools in an upcoming entry.


In this entry I’ve covered my annotation workflow and how I make use of my annotations when writing. Using Dropbox, I can access my highlighted manuscripts from anywhere. Using a combination of Papers and Devonthink I can make use of my annotations saving me countless hours of time when writing.

Tagged with →  

14 Responses to Writing Workflows: Capturing Annotations to Enhance Scientific Writing and Knowledge Retrieval

  1. Hi Jeff, thanks for sharing your annotations workflow. I was wondering if you can provide more details regarding how you use Devonthink. I use Papers 2 to manage articles and notes, then I use Bookends to manage my citations references and then I go directly into Scrivener. What are the benefits of DevonThink? Thanks for your help.

    • Thanks for your note. Devonthink has “Artificial Intelligence,” although that’s a bit of a misnomer. The AI in DTP interprets a search and matches notes that wouldn’t be found if you use only Papers. I use DTP for all sorts of things (it’s well worth the cost of the Pro Office edition).

      I plan a series of future entries on how I use Devonthink–please stay tuned!

  2. Alex says:

    Great post! Your method of getting Papers2 data into DTP looks nice and clean, although maybe a bit cumbersome since you need to copy and paste each note.

    Someone has written a script that does this whole process and allows importation of individual notes and referencing. Check it out. It may streamline your approach.

    Question re: indexing. When you index, does DTP access the notations and highlights you’ve made within papers? If so, why then should we import the actual RTF notes into DTP? I’m still a newbie DTP user, so forgive me if this is a silly question.

    Keep up the great work!

    • Separating each note is a critical part of my workflow. I previously tried the script you mention above without success. I will try again.

      The reason I separate each comment and note into it’s own file has to do with avoiding information overload when searching / writing. By separating each comment, and putting each side-by-side with the annotated manuscript, I avoid having to read through the entire paper (or pages of notes) to find a concept–Devonthink does it for me. Devonthink returns BOTH the annotated paper AND the individual comments relevant to my search. I then cut and paste each individual comment directly onto a “concept card” in Scrivener without having to search further in the original manuscript.

      • Alex says:

        Hi Jeff,

        I hear you about the individual notes. The script has been updated, and the latest version allows the import of individual notes from the single body of notes in papers2. If you use a colon (:) after each note in papers2, the script recognizes it as an individual note and imports it into DTP as its own RTF file. It comes with the magic manuscript citation in the RTF too, so you don’t have to do that manually.

        I’m really looking forward to seeing a future post about your writing workflow with DTP and scrivener. I’d love to see how you use DTP to search your papers2 notes, and then how you use those build a manuscript with scrivener. I think the best aspect of scrivener is the ability to manipulate the text using the corkboard or outline. My experimental approach has been to make DTP hyperlinks to each card in scrivener, compiling all the references I will use to make a paragraph or argument. I haven’t really done too much writing yet using this method, so I’m open to hearing about better ways.

        Keep up the great work!

        • After you posted your note, I went back and tried to run the script again. Because I don’t keep my Papers2 database in the default location, I had to alter the script a bit. Even after doing this, the script didn’t work for me. Another negative, it is written only to export NOTES of a manuscript (not highlights).

          I was, however, able to use the Skim script to parse highlights and notes. This will be a huge time-saver moving forward.

  3. Alex says:


    Sorry about your trouble with the papers script. You’re right about the highlights, which is indeed a limitation. Did you modify that Skim script to work on papers2? If so that’s super useful. Would you mind sharing that?

    • I didn’t have to modify it–just be sure the paper you want to process is open in Skim. Unfortunately, Skim can’t read Papers2 highlights, they have to be converted to Skim Notes–so it won’t be possible to batch process your library, at least yet.

      See today’s entry on how I accomplish this.

  4. Joe Bianco says:

    Hi Jeff:
    Just out of curiosity, how do you get highlights from an article to auto populate in a notes pane on the ipad? I really want to be able to have all of my highlights collected in a single note automatically, but can’t figure out how that works on the iPad.

    • I typically don’t use this feature. I highlight and make notes using the highlighting tool in Papers. I think you have to write directly in the notes field on the iPad for the app to log your notes.

  5. Dellu says:

    Great workflow. I do have exactly the same workflow. But, I use Sente instead of Papers 2. A script exports individual notes, quotes and comments from sente to DT (Scrivener).

  6. Neal says:

    The link for the script from Papers2 to DevonThink no longer works. Does anyone have a new link?

  7. Matthias says:

    Hi Nea, for Papers 3l, I’ve written a script that exports all notes & highlight annotations of all selected publications to DEVONthink.

Leave a Reply

%d bloggers like this: