phoobar.org

Research with DEVONthink, Zotero and Dropbox

June 3, 2013 written by Nikolaj Raahauge in 2212 words
DEVONTHINKPRO ZOTERO RESEARCH DROPBOX IANNOTATE SKIM PRODUCTIVITY
Sean Connery doing research!

To be honest, I find it hard to identify the relevant stuff, zapping through different folders, which always gets messed up, with a lot of duplicates and forgotten documents as a result.

As part of my work I often read a lot of articles (typically .pdf’s of blog posts) which I use as inspiration for workshops, presentations, white papers and other assignments.

I most often use my iPad for reading and annotating with Dropbox as storage for the files. I also depend on a bibliography manager for citations and keeping track of authors, publication date etc.

When it comes to the actual writing, I like to toy around with the outline to get the disposition right, pull up an editor and write in fragments, drawing on notes from the various sources I’ve collected.

In the coming sections I’ve described my considerations and choices for my own research & writing workflow.

Information Management System

I started looking for an app that could help bring more coherence to my information management, and stumbled on an essay and a blog post respectively, by the author Steven Berlin Johnson1.

In these he described his needs and requirements for information management along with his writing process. He also explains how he uses an app for Mac called DEVONthink to help him with the management of all the information that his research consists of.

Mind you, I am in no way near the complexity level or requirements of the good mr. Johnson, but the simple way it was possible to organize data combined with the search features of DEVONthink Pro (DTP) appeared very attractive.

I downloaded the trial edition and started poking around. I’ve read several places that DTP has a very steep learning curve and it does take a while to get used to the screens and basic functions, but a lot of the difficulty comes from organizing the structure of groups; it’s quite free form in that respect.

When that is over though, I really started to enjoy the simplicity of the different screens. Not that modern looking, granted, but very functional and easy to use. Some of the features that I quickly started to enjoy were:

The search function When content is added to DTP it scans and indexes the text, making all of it search-able. The search functionality is the best I’ve experienced in this type of app, having many clever features like the x NEAR\5 y function that will match y within 5 words of x.

See also This function is particularly useful since it lists items related to the search result, ranked by relevant score. For example, after performing a search on agile NEAR\5 culture, I can select any of the results, click on the See Also hat and get a list of hits that are related. This feature is nearly invaluable because it actually facilitates serendipity2; stumbling over pieces of knowledge you didn’t even remember having or knowing. This is exactly what I meant by needing more coherence between all my bits of information!

Auto Classify Once you start constructing and populating your database structure with groups and files, DTP’s Auto Classify feature will get stronger and stronger. Basically, once a file has been added to the inbox of the relevant database, DTP can use it’s AI to place files in the folders of similar material. This, of course, is very dependent on the group structure you’ve arranged to be successful. The more explicit it is, the better it will work.

Indexing files This one might not be that obvious, but DTP can handle files in two different ways:

  • By importing the files to the database file structure3
  • By indexing files from a location on disk, without adding them to the database file structure.

In both cases, DTP will scan the files and index the contents. This way it is possible to add e.g. a folder from Dropbox and it’s contents to DTP. The contents will then be indexed and become part of the database, but the files remain in the Dropbox folder. One rather annoying thing is that the files must be refreshed manually4.

Indexed Articles Group from Dropbox in DEVONthink

So, how does it actually help me? Let me answer this by listing some of the most frequent tasks I perform in DTP:

  • Researching a certain topic will make me start to collect loads of different material from several different sources: blog posts, web articles, images and videos. I then start to identify relations between the different sources (and this is where the AI really comes in handy) creating a group structure that can be developed as I go along. From then on, what will actually happen with the collected material depends on what I’m researching for, if it’s a presentation, a white paper or a blog post.
  • Receiving e-mails with attachments will make me use one of the supplied OSX System services from DTP that will automatically import the e-mail along with the attachments to the Inbox of the default database (or the Global Inbox depending on your preferences.
  • Keeping track of letters, invoices, printouts from the bank or even iMessages or other log type information.
  • Bookmarking various information is now something I do with DTP. It’ ability to make snapshots of websites and save them as .pdf is especially useful. It can also double as a RSS reader, which makes for quick import to the database.

I ended up purchasing a license for the Pro version as I found the extra features of the Office version to be more than I needed, but it’s available for trial as well. My recommendation is to try out both versions to find out what meets your individual needs5.

Reference management

Another part of my information management requirements was to establish a bibliography of the books and articles I have lying around, using a Reference Manager application. New stuff comes in every week from many sources, especially after I discovered the articles on Harvard Business Reviews website and got a good Google Reader setup. Other sources are blog posts or web-based articles and so a requirement was that they could be added and properly cited.

This led me to try out many different bibliography management apps for Mac OSX, from Papers2, Bookends and Sente to Zotero, Mendeley and Bibdesk.
I finally settled on Zotero, which is free and has some specific features that I find really awesome:

RTF Scan Zotero scans the .rtf document for citation references like {Boehm, 2004} and matches it with the relevant entry in its library. You can choose the style you want your bibliography to use (Zotero supports a lot) and the scan will then insert the matching bibliography either as notes on each page or on the last page of the document.

Microsoft Word plugin The Word plugin enables insertion of citations and a bibliography from a pop-up, which let’s you search directly in the Zotero library, listing matching entries. This is pretty neat if there are several publications from the same author. The only caveat seems to be, that if the file already has a Zotero bibliography created by the RTF Scan function, it will not respect it and thus it will get messed up (unless I’ve misunderstood something, which is entirely possible).

Dropbox integration through the Zotfile plugin Zotfile enables me to simply keep the bibliographic information in Zotero, along with a link to the file in its Dropbox folder. The plugin also has other cool features such as automagically renaming files etc. using the bibliographic data. As a consequence, my files are not synched to the Zotero service either.

For me, the runner-up was Papers2, but with the amount of features it had over Zotero compared to the price difference, I decided it wasn’t worth it. Of course, Papers2 does have an iPad app, which it can sync with, but since I’ve put all my articles in a Dropbox folder (and I find that both flexible and awesome) I do not need an app to sync with. It’s really cool that all I have to do, from any device or computer, is to add a link, a .pdf article or document to my Incoming Articles Dropbox folder, and when the time is right, I boot up Zotero and add it to my library which will then add it correctly to my folder structure using the Zotfile plugin6.

UPDATE: I’ve since started using the Editorial app for iPad by omz:software, when writing ‘on the go’. I had to find a way for my references to be available, and for that I created a simple workflow which you can read about in this post: Citation Workflow for Editorial.

Reading and processing

Now, as I’ve mentioned, quite a lot of the research stuff I gather are blog posts or .pdf articles; that’s all well and good, but it’s not enough to just manage it. For me, gaining knowledge is a process that requires more than just reading the material. I have to take notes, and I have to siphon through and comment my notes as well, for much of it to stick.

What tends to work the best for me is a combination of different study methods. So, when I go at it with a text that I know I must understand, process and be able to integrate in other things, my process is usually like this:

  • What is the context and what would I like to use the knowledge for?
  • To what depth do I need to know the material?
  • I go through the text using the SQ3R method while taking notes during the Read step.
  • I consolidate notes and annotations in either nvALT or as an .rtf file in DTP, depending on the source. Typically, nvALT stuff are for the smaller parts.

My favorite apps for reading .pdfs are iAnnotate for iOS and Skim for OSX respectively.

My reading workflow incorporates several tools. What I’ll go through here is the flow when I need to read and understand .pdf’s in-depth:

  • I’ll start out with adding the .pdf to Zotero which adds it to my Articles Dropbox folder structure.
  • From here on it is indexed by DTP and becomes part of it’s database.
  • Once the file is in Dropbox I can open it in either Skim or iAnnotate, depending on the device I’m reading on (iPad or MBP). I then do my highlights while reading and attach notes as well.
  • When I’m done reading the file I will open it in Skim and run an Apple Script created by Rob Trew7 which will import the notes from the .pdf file as individual .rtf notes in DTP. This makes every annotation part of the AI searches in DTP and is probably one of the most powerful aspects of my workflow.
  • I then merge the .rtf files, which will add a new file to the DTP folder structure with the individual notes in it. I go through the notes in their entirety, adding comments or links to other important pieces of information.
  • When I’m done with the notes file, I move it to the Dropbox folder of the .pdf (along with the individual notes). This gives me the .pdf file, with annotations, along with the individual notes and a consolidatet, abridged summary of the notes.

Now, when It’s just smaller things like individual blog posts, articles or other, my notes are typically thrown into nvALT, which again stores notes in a Dropbox folder, which again is indexed by DTP.

So, this is my workflow for information management and research using:

My next blog post in this vein will cover larger writing pieces in Scrivener together with Zotero and DEVONthink.

Thanks for reading what got to be a really long post! If any questions arise, don’t hesitate to contact me on twitter @ncraadk. Peace out!

  1. SBJ’s Where Good Ideas Come From was an awesome read, it covers serendipity as well as other important factors for innovation to thrive, in great detail.

  2. This word and the concept behind it, was new to me. But it actually describes what happens perfectly: Serendipity means a “happy accident” or “pleasant surprise”; specifically, the accident of finding something good or useful while not specifically searching for it.

  3. DTP keeps imported files in its own file structure, with a folder hierarchy.

  4. This is easiest done by selecting the top folders and hitting ⌘-⌥-S. Of course, you might have many folders to sync, which is why it isn’t always attractive to have stuff like that act on it’s own since it could be a heavy process.

  5. The Pro and Office versions both offer the possibility of using more than one database, which is particularly useful if you plan to use it for large amounts of data, the smaller the database, the more focused the AI will be.

  6. Wikipedia has a nice page which compares aspects of a multitude of different reference/bibliography managers, you can find it here.

  7. You can find the script here along with many other useful DTP scripts.