Dating DEVONThink

I was all excited. I thought I had found a solution to my vexing Journler problems. Crashes, freezes, all manner of frustrations had pushed me away from the love of my life application. Journler was the app I had been immersed in for the first year of my writing, my first crush, but the application’s solo developer had gone off, and it was clear, the bugs that existed would be problems forever. So I went looking for another.

I wrote “Bye, bye, Journler. DEVONThink is my girl now. She’s not beautiful like you, kind of clumsy actually, but she is smart and will be here for the long haul…”

Now I’m starting to wonder. DEVONThink is an application best described as inattentive to its appearance. It is messy in the way it interacts with others and is more worried about the mad scientist, artificial intelligence core of the program than in adding any real value to how users create or manage data.

A web buddy of mine, who also searched for another after her Journler romance failed, called DEVONThink “realy, um…. German.” She was right. DT is rigid in its requirements and often sports a bad haircut. You get the idea.

But I’ve got to say that DEVONThink can hold her beer. The program runs and runs even with huge data sets. One user, a DT Evangelist, has over 150,000 journal articles in his DT folder. Another dumps dictionaries in for reference. I’ve realized from the multiple windows that DT offers, each a bit more quirky than the last, that I will need that kind of power, because I can see for the first time just how much stuff I’ve accumulated.

It was this Arian capacity for endurance that seduced me to DT.

My sense is that once you begin using info manager tools, which is the loose category both these programs fall in, there is a linearity, perhaps an exponentially, to the data you collect. Having access to a library is useful, but having your own library in a tool is totally cool, so you start to capture and produce more and more and more stuff. Steven Johnston, who wrote a ‘famous for the web’ article in the New York Times about how he used DT to write his books, called these applications groundbreaking in how they transform the creative process. I agree.

But the better question is what do these apps do for you that the file system doesn’t? If there is no work being done by the application then it is just an additional layer of code that will eventual corrupt or become obsolete (Anyone here ever use Ecco Pro back in the 90’s? I bet your data is lost too.) That was the point recently made by Alex Payne who said, dump the apps and just use OS X’s file structure. I’m not sure I agree, because at their best these apps enhance the operations of the file system, in other words they do things for you.

Like tagging for instance. I’ve never been exposed to a better system for dealing with multiply interrelated data sets than tagging. For example: A simple text file recording daily activities can include notes on an art exhibit, a snip about the cute girl at the front desk, thoughts about the economics of the art market, the address of the exhibition catalogue publisher, and an idea for new shelves it the closet. With tags that item can be referenced many ways based on the content needs at the time. An application that helps you will add tags when you add an item to it, it will allow for predefined searching and combination of tags, it will make editing tags in one or multiple files easy. DEVONThink does none of this. The evangelist with 150,000 items claims he’d rather use the DT’s artificial intelligence to search than tag, so you should too.

In fact DEVONThink doesn’t really do much. It doesn’t have a common view of multiple folders. It is more cumbersome to use than directly editing Spotlight comments. The text editor is forlorn. For some bizarre reason shortcuts (which they call replicas) are always labeled red. Wiki links do not open in their own window. Tabs look like they are from the 1990’s, refuse to open in a separate window, and resist any reorganization. Labels are written to exported files but are not included in imports. Dozens of keystrokes have to be modified to match standard OS X functionality. Labels sort alphabetically by label name. File metadata is not accessible or searchable, and dates are considered sacrosanct, not editable except by a ‘who knows if it will be there tomorrow’ script.

And that’s how DT solves most of its interface problems. It gets users to write scripts. That’s sloppy programing. What you end up with is a file system on top of the file system with some wiz bang search capabilities, running cobbled together scripts built by users. But it’s not one that helps you very much in getting your work done. The scripts just help overcome the deficiencies of an introverted application. And that’s too bad, because I have a feeling that I’m going to be stuck in DT for a while. There really isn’t anything out there to compete with it.

Journler was the answer, but as I’ve just learned, unfunded solo developers are very risky. Who else to turn to? Together concerns me because it gags on data sets over a few thousand items. Evernote’s business model is committed to datalocking you into the application, which is intolerable. Yohimbo is a starter tool not a long haul application for a serous user. Lastly the Lit / Citation managers, like Papers, Zotero, Endnote and Sente are indexers, not writer’s base pads.

So yea, DEVONThink. She’s good right now, but I’m lookin’

(Note: Yes, yes, I know that DT 2.0 is still in beta and that improvements are coming, which is kind of like saying “She’s off to charm school to take care of that nasty habit of drooling at the restaurant”. Let’s see if the developers, who I grant you are very customer focused, will realize that the future is the front end of usability not the back end of data searching – DDB)

More on file systems, notes and archiving on Dougist…

Shifting Mediums

Writing Tools – Journler

WriteRoom and Notational Velocity

The Low Fi Manifesto – Data Architecture, and Journler

I’ve moved on … see here >  File System Infobase Manager

This entry was posted in Best Of, Productivity and tagged , , , , , , , , , , . Bookmark the permalink. Post a comment or leave a trackback: Trackback URL.


  1. Posted February 21, 09 at 12:49 pm | Permalink

    Hey Doug: you’ve tried Scrivener already, right? I was a long time Journler user, but started to get lost in my own collection of items. I’d been using Scrivener to organise my writing, and decided to shift project notes over there as well. Does me fine so far. Of course, you have different requirements from me, so ymmv…

  2. Doug
    Posted February 21, 09 at 8:13 pm | Permalink

    Hey Jacob,

    I write in Scrivener. What a great application. I’ve heard of people who use it organize their notes as well as write. As for now I put notes into Scrivener that I “pull” from my info-manager, whatever that might be at the time. I assume that notes in Scrivener are copies and therefor can be trashed. THat way I focus on output with Scrivener.


  3. Posted February 25, 09 at 10:55 pm | Permalink

    I’m with you on this one, and I still haven’t totally gotten my head around DEVONthink. But I think we have different reasons for our dislike.

    Tools that hide and torture file hierarchies while totally depending on tags bother me. I reluctantly use iPhoto and iTunes because I finally gave up resisting. But that doesn’t make it any easier to move into a tool like DEVONthink.

  4. Posted March 12, 09 at 8:00 pm | Permalink

    Hi Doug, I too just fell in love with Journler and over the past two weeks have begun to dedicate myself to it completely. I am a writer and an academic and this investment is dire – being ripped apart from my hard work by the non-support of a proprietary format (which for the most part journler is not) is unthinkable!

    How did you get your Journler stuff into DevonThink? A lot of people are worried about the Journler developer dropping off lately, and a description of your transition process would no doubt be a godsend to many. Thank you for your time!

  5. Doug
    Posted March 17, 09 at 9:19 am | Permalink

    Hi Caneel,

    Thanks for your post.

    It looks like Phil Dow (the Journler developer) is back, at least for a while, at least on the Journler boards. He seems to be actively working on the next release of the application. But even so, I think this whole experience has put me in a different place with regards to “shovel box” programs, the term of derision used by Alex Payne to describe these apps.

    While I found great value in using Journler to create a lot of writing work, the vector of my development and production began to outstrip the application’s functionality. These tools really are just file system replacements that add a few functions, like…

    1. Tagging

    2. Immediate editing

    3. Browsing

    4. Advanced search

    5. Wiki like links

    While I miss (1) (Spotlight comments are fraught with potential problems as has been documented widely, and may now be replaced by openmeta tags), I found easy solutions for (2) and (3) using Bean and Quicklook. (pushing the spacebar really isn’t that hard) Someday I’ll have to deal with (4), but for now DEVONThink will allow me to use it’s searching capability against my OS X file system should I ever need ‘more than Spotlight can give me’ capability. (5) is replaceable with Shift CMD F, and you get a better set of results on common terms (mark any phrase you would have wiki linked in your rtf /d and Shift CMD F will find all like phrases on your Mac…cool)

    I’lll also add that I‘ve learned that Scrivener, which I write my bigger works in, has a lot more functionality than just drafting and writing. It really is a fabulous binder tool and can replace many of the draft management functions for which I used Journler.

    The real value I gained from Journler and the DT conversion was the process of architecting my data. They made me think about my folder structure differently and eased the process of moving everything around. I don’t think I would have done that as easily without them. Ironically they also made it easy for me to leave the applications behind.

    I wrote about moving from Journler to DT here

    A few more comments…

    It was a pain. If you’ve imported your resources (as many academics do with their journal articles) then you’ve got to fish them out once you import everything into DT, or else you have a whole series of silly one or two item folder trees.

    You loose functionality. The beauty of Journler was the idea that I could take notes ‘as if on a piece of paper on top of my pdf’s’. Like we used to do before all this fancy technology came along. No other system allows that. They all make you have your notes as a rtf/d ‘next to’ your research or writing. As an example, I added each draft of my fiction work as a resource to Journler and then, above or below the link in the entry I would write notes like “Dr so and so says this is great,” or “Publisher says, this draft will never sell with out a sex scene,” or whatever.

    Moving out of DT is a breeze. DT doesn’t really do anything to your data. So once you’ve edited out the Journler folder structure (manually), you just dump exports and that’s that.

    You gain freedom and scope. My data set (because that is what it becomes over time) had outgrown the tools. I was constraining them into the applications. Now they are just there. If an application can work with it there, great.

    The Alex Payne article I cited above really shook me up. He went off on these apps as a category, and after thinking alot about that, and ratliff’s post on the Journler forums , I don’ think I’d ever go back to using any of these applications the same way again. I now have all my files outside of any application and I’ve found that given OS X’s basic functionality I can manage and create just fine, perhaps better than I could in DT or J.


  6. LexM
    Posted April 23, 09 at 3:58 pm | Permalink

    Just found this by way of the Journler forums. With respect to the Alex Payne post, have you looked at Tinderbox developer Mark Bernstein’s thoughts on Payne’s assault on everything buckets? Bernstein writes (this is an excerpt; full post at

    “Payne’s main point is that applications create and use structure. They are better than buckets: trying to do everything in a single application leaves you doing lots of things with a versatile but second rate tool.

    What Payne misses — what nearly everyone has missed in thinking about the question — is the process of finding and creating structure. Yes: you want to keep things organized. Yes: you want specialized tools for special tasks. But things don’t arrive with structure (and, when they do, they have the wrong structure!) and the kinds of structure you want are always changing.

    So, Payne is wrong: you do need an everything bucket because sometimes you get a receipt or a podcast and you just want to say

    ‘Keep this, don’t lose it, but I’m really busy now so GET IT OUT OF MY FACE.’

    And sometimes you want to say

    ‘I have no idea what this means, but I bet it will make sense after I learn to read Farsi.’

    So, sure, keep an inbox, or a bucket. Use the file system if you like, use Yojimbo or Devon or whatever if you like. Use ’em all: you can have two separate piles of paper in your office!”

    I take Bernstein to be saying that Payne’s approach is what gave rise to everything buckets in the first place – you have to know what structure everything belongs to before you put it into the file system or before you put it into the special purpose application, and life’s not (always) like that. Payne is telling you to impose structure from the top down; Bernstein is saying that too often structure has to emerge over time, from the bottom up, and until it does unstructured data has to reside someplace(s).

    I don’t have a dog in this fight, but my own workflow employs an everything bucket (MacJournal, which I moved to not altogether happily from Journler [for the same reasons many of us have emigrated from Journler], which I moved to when Circus Ponies NoteBook got over-complicated and under-stable, which I moved to from OmniOutliner when…you get the idea) where things sit until they become Scrivener projects, emails, or…whatever they might become.

    Perhaps Bernstein is just saying things that make me feel good about my ongoing state of semi-organization (or things that make his Tinderbox look like the ideal information management tool). Whatever the case, this an interesting discussion.

  7. Posted April 24, 09 at 11:14 am | Permalink

    did you ever have a look to EagleFiler ? I supports the OS filer system. I use it as the »Everything Bucket«.

  8. Posted October 13, 09 at 5:42 pm | Permalink

    I found this blog very helpful. I have tried to like DEVONthink, but it is just so clunky and it does not integrate very well with Spotlight. Fortunately beauty is not everything, because DEVONthink seems extremely powerful. Yet, I prefer the elegance of Evernote. Sometimes I wished I could take the two and merge them together.

    I will keep an eye on your blog because I would love to hear more about what you think on this topic. Yours is one of the best blogs I have seen to date.

One Trackback

Post a Comment

Your email is never published nor shared. Required fields are marked *

You may use these HTML tags and attributes <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>