On several occasions I have written about how I use Devonthink Pro (DTP) for scholarly writing. Although at one point I had all my information on manuscripts in a single DTP database, over the last year I have maintained separate databases; one for PDFs, another for my annotations.

I accomplish this by having separate folders in Dropbox. After extracting my annotations from Highlights.app, I place the exported files in their own folder. Once a month, I export all the PDFs in my Paper’s Library to their own Dropbox folder. I use a Hazel script to throw away any duplicate PDFs in the Dropbox folder. I index (not import) the annotations folder into one DTP database, and index my PDFs into another.

This setup allows a fair amount of flexibility. Not only is this setup advantageous for writing with DTP (as I will cover in my next entry), it allows easy access to my PDFs for reading with Liquidtext or listening with Voice Dream.

Tagged with →  

2 Responses to Folder Structure to Maximize Writing with Devonthink Pro

  1. rickla says:

    On the surface, this would seem to be a retrograde step, in the sense that removing the PDFs from the database with annotations removes the possibility of selecting a key passage within one of your annotations and having DT find a passage (through See Related Text) in another PDF that you haven’t yet included in an annotation.

    But it seems to be of a piece with your use of Tinderbox ( http://wippp.com/blog/sense-making-academic-literature-using-tinderbox/ ): the way I interpret what you are doing is that you are trying to get greater manual control over the information that presents itself to you, generally with a view to reducing that information rather than increasing it. It seems to me that DevonThink untamed (i.e., a single database fed with all the PDFs you’ve ever come across that could possibly be useful) overwhelms you, even though it’s possible to use Smart Groups (or tags if you want to be more manual) to drill down to the subset of information you want.

    Apologies if I’m reading my own thoughts into what you are doing but the things you have said recently, both here and on, e.g., the Tinderbox forum, have chimed with my own vague feelings of unease or discomfort with the ways I have been working with DevonThink: its functionality makes me feel that it should fill all my needs between(non-inclusive) storing bibliographic information and actual writing, but somehow I never feel all that comfortable with it. I’ve ceased keeping any text in my main database that hasn’t already been read and selected as relevant by me, and I’ve started experimenting with Tinderbox, based on the system that you described in the post linked earlier.

    Thanks for continuing to share your ideas and discoveries, and I’m looking forward to your next post.

    • You are right, I’m trying to get more granular control of the information in my library–both read and unread. I have taken to extracting entire manuscripts I’m reading and those I’ve written into 1-5 paragraph chunks (sometime with comments, sometimes without) then storing them in their own DropBox Folder. I do the same with grants. I keep ALL my PDFs (read and unread) in their own DropBox Folder indexed to its own database. I tend to use sentences when searching in DT. Querying my annotations database brings up smaller chunks of information I’ve read (e.g. the salient paragraph in a 10 page paper). I use the same technique on the PDF folder to find the manuscripts that have information similar to what I’m writing as a way of focusing on what I should read next. I’ve found the separate folder structure gives me more flexibility (e.g. the ability to point apps like LiquidText and VoiceDreams to the folder) and returns less chaotic results than I would have if everything was mixed together.

      Although there is overlap between what I do with DT and Tinderbox, I find them useful in different ways. I like the visual organization of Tinderbox. As an example, I’ve been collecting information on the human microbiome. This area is developing so rapidly, it’s very difficult to keep up. I’m feeding abstracts (primarily of unread articles) into Tinderbox to give me a visual / outline grasp of the domain. When writing, I use Tinderbox very early on to focus my attention on things I have not (but should) read from a 30,000 foot level. I use DT to get down into the weeds.

      I could (and have started) replicating my Tinderbox queries with DT Smart Queries–but I, like you, find the amount of information DT returns to be daunting and unorganized. That being said, it is a fair amount of work to extract the abstracts to Tinderbox — I’ve looked for more time efficient ways to accomplish the same thing–so far no luck. As I’m sure you can tell from my musings over the past several years, this is all very much a work in process.

      I appreciate the thoughtful comments.


Leave a Reply

%d bloggers like this: