I first noticed the buzz surrounding DEVONthink way back when I was putting together my "Outboard Brains for OS X" article back in 2003.
Several people mentioned it as a worthy kind of snippet archiver, one that offered unique features. I did try it out, but like other newcomers to the software, my first attempt left me feeling a little nonplussed. I couldn't see the usefulness.
Then earlier this year, author Steven Berlin Johnson wrote a widely linked and oft-cited weblog post about how he made use of DEVONthink as a reading and research tool. By throwing enormous amounts of data into it, he was able to make best possible use of the application's built-in semantic searching and cross-referencing. The result was that DEVONthink made suggestions and links between materials that Johnson could never have spotted himself; it actually did some of his research for him.
This kind of endorsement sparked a lot of interest in the software. The buzz really took off, and over time it metamorphosed into speculation as to what shape the next version of DEVONthink would take.
Now the world of productivity geeks is turning a thoughtful eye toward Devon Technologies, following the release of DEVONthink Professional 1.0.
For a couple of weeks, a beta version of this application was available as a free trial, something that has prompted a lot of people to download it and experiment. Initial reports were varied, but some users said the new features and Tiger-specific technologies make it a very exciting release.
So here, for the uninitiated and the unenlightened, we present a guided tour of DEVONthink Professional. Our guide is based on the beta version, so some features may have been changed in the final release. What does this app offer and how does it work? More importantly, why are some people so buzzed up about it? Come with me, and I'll show you.
One thing that can leave people flummoxed when they first use DEVONthink is that, on starting it for the first time, there's not much to see. The standard view offers a toolbar at the top, a folder hierarchy on the left, and a large viewing/editing window dominating the screen--which on initial start is empty.
The best way to think of DEVONthink is as a database for the rest of your stuff. It's empty because, right now, it doesn't know where your stuff is. You just need to tell it.
The simplest way to do this is to tell DEVONthink to import your documents; just drag your Documents folder into the window and sit back while the application builds a database of all of your information. DEVONthink indexes everything you throw into it, building up a map of the content of the files and their metadata.
But as well as being a database environment for viewing what you've already created, DEVONthink makes a very good editing environment for creating new stuff. With built-in RTF, plain text, and HTML editing smarts, it becomes a very capable and helpful writing tool. Best of all, it makes connections between what you're writing and the rest of your database, while you are writing.
One of DEVONthink Pro's most important new features is the ability to manage multiple databases (in previous versions, everything had to be stored inside of just one). Unfortunately, only one of those databases can be open at any time. Should you wish to switch to another, DEVONthink saves everything in the first one and closes it first. This might be frustrating if you often intend to move data from one database to another.
While DEVONthink's viewing panes can be rearranged in many ways and dragged around to suit your monitor settings and the way your eyes work, the simple concept is that the app always offers a view of the files currently being used, searched, or indexed, and a view of the one file that is the current focus of your attention.
There's one further thing, a feature unique to DEVONthink. At the bottom of the window are two buttons marked See Also and Classify. These are essential to DEVONthink's success as a personal database, since they are the controls used to find, or create, connections between different documents. We shall look at them in more detail later in this article.
Once you've dragged some documents into DEVONthink, it's time to start exploring what it can do.
To start with, your data has some degree of built-in structure, especially if you have imported files from the Finder.
But to get the most from DEVONthink, you need to add further structure, making links and connections between files and groups of files.
When DEVONthink talks about "groups," it means documents that are bunched together because they share some meaning or theme. You can create groups yourself, and simply drag documents into them. Groups can have sub-groups, and they appear in the DEVONthink GUI as folders, so it helps to think of them as DEVONthink's Finder alternative.
Throughout DEVONthink, in search results windows and document windows, you will see the Classify button. Click it when viewing or previewing a document, and a drawer opens, offering you likely groups to which your document belongs.
Note that DEVONthink has already picked these for you, and is making suggestions based on the existing groups' contents. A brightly colored line icon appears next to each group suggestion; the longer the line, the better suited the current document is to that group.
What's happening is DEVONthink is trying to help you. The job of classifying documents remains yours, but the software attempts to make it quicker and easier by spotting themes and commonalities between documents and anticipating how they might be grouped together.
To move a document to a group, choose one of the suggestions from the list in the drawer and click the Move button. The behavior is like that of the Finder; the document disappears from where it was (either in another group, or ungrouped) and shifts location to the new group.
Remember that all of this is happening within DEVONthink's database; your original files, the ones you imported in the first place, are not touched.
Of course, if you have a lot of data (and as we shall see in the conclusions later on, DEVONthink is the kind of tool that suits people with a lot of data), the thought of manually classifying thousands of documents might be a little off-putting.
So there's an Auto Group command under the Data menu, which does a surprisingly good job of collating a mess of disorganized files into some kind of order. It could save you hours of work.
Documents can belong to more than one group. Right- or
Control-click a document, and you'll see a "Replicate to..." option. This makes an identical copy of your document (not an alias) and puts it in the group of your choice. The replicant document mirrors the original, even if the original is changed, and vice versa. Replicants are synchronized copies of their originals.
Duplicates, however, behave differently. A duplicate document exists independently of the original; edits made to it will not be reflected back.
You can ungroup things, too. Ungrouped documents move up one level of the groups hierarchy, all the way up to the top if need be, and will stay there until you assign them to a group.
There's a subtle difference between "grouping" and "classifying" in DEVONthink. Grouping is the process of creating groups of documents; classifying is the task of assigning documents to existing groups. You are more likely to spend time grouping when you first start using DEVONthink, and classifying once you've been using it for some time.
Also, DEVONthink works better if documents and groups are kept separate. In other words, don't have documents sitting alongside groups within the hierarchy; instead, create a new sibling group and put the spare documents in there.
Your data will be displayed as a hierarchy of folders, just like the Finder. If you've dragged in your entire Documents directory, its structure will be faithfully recreated within DEVONthink.
DEVONthink is very good at finding things within this large pile of information. The quick find box in the toolbar offers a variety of search options and can zip through even large databases with impressive speed.
F brings up a separate search window that offers much more power. Enter a search term; but before starting the search, try clicking the Spelling or Context buttons. These open a drawer with a list of additional search terms that DEVONthink considers similar, either by the way they are spelled compared to your search term, or by the context in which they appear within the database.
Double-click one of these suggestions, and the search is done on that term instead. Thanks to the way the search window is designed, you can view the document you find before opening it.
Another important way of finding related content is the See Also button that appears when viewing any document.
This really comes into its own once you have spent some time organizing your groups and classifying your documents appropriately.
Opening the See Also drawer brings up a list of documents that DEVONthink considers similar to the one you are viewing now. Select something from the list, and you can read that in situ, while you still have the previous document in your head. Now you can perform a See Also search on this second document, repeating the process; this is exactly what Steven Berlin Johnson described doing in his influential weblog post.
He concluded that having small, granular snippets of information was much more useful than large documents: "I have pre-filtered the results by selecting quotes that interest me, and by archiving my own prose. The signal-to-noise ratio is so high because I've eliminated 99 percent of the noise on my own."
Which doesn't mean that you can't, or shouldn't, have long documents in your database; but Johnson's identification of the 50-500 word count "sweet spot" is probably one of the best tips you can find for long-term usage of this application.
DEVONthink Pro can count the words in a document, but in more ways than one.
Sure, it will tell you how many paragraphs, words, and characters each document has, but the analysis of words goes much deeper than that.
Much of DEVONthink's thinking is done by examining concordance, or the number of times each word appears in each document.
You can examine any document's concordance with Tools -> Concordance, and detailed word count stats will appear:
From the chart, you can hit the Search button to show other files that contain the same word (great for hunting down obscure links), or the Similar button to show words with similar meanings or contexts.
Control-click a document, and choose State -> Show. The document's icon will be replaced with a checkbox. Standard OPML files can be imported and exported.
Control-click, and choose Set As Title; the document's name will be changed accordingly.
), allows you to grab text from any other Services-aware application and instantly create a new note in DEVONthink.
DEVONthink is a big tool for big tasks. If you need a simple writing tool and your needs are basic, then TextEdit and Spotlight, and perhaps DEVONnote, are better choices. The same applies if you're looking for an outliner or an RSS aggregator.
DEVONthink Pro is for people who have a huge amount of information that needs to be cross-referenced, indexed, and sliced and diced in ways that the user might not have considered before.
If you collect any sort of text-based data or write a lot of text, and are prepared to invest some time in getting it into DEVONthink in the first place, the rewards can be significant.
This application is ideally suited to a wide range of people who handle lots of data but find it hard to keep an overview of it in mind. People like students, authors, academics, or project and people managers.
The biggest barrier to adopting DEVONthink as your information management tool is the initial task of importing, grouping, and classifying; if your data store is already huge, this will be daunting and possibly frustrating, even with the use of the automated grouping and classifying tools.
But this application is one that rewards your investment of time. If you use it often, to organize lots of information over a period of time, you will soon discover the benefits: powerful searching and cross-referencing, and an ability to make semantic, intelligent links between what were previously unsorted and unhelpful data.
Giles Turnbull is a freelance writer and editor. He has been writing on and about the Internet since 1997. He has a web site at http://gilest.org.
Return to the Mac DevCenter
Copyright © 2009 O'Reilly Media, Inc.