Using Smart Groups

I’ve recently gotten way back into DEVONthink as a means to capture, wrangle, and analyze all the reference material in my world.

If you’re new to this amazing application — and at the risk of far exceeding my understanding of both the human brain and this particular piece of software — DEVONthink learns the neural pathways between the stuff you know or say is related. But, more importantly, it prompts you on the relationships you probably don’t know exist (yet). This is awfully useful and wildly stimulating to the busy front parts of my own brain, such as it is.

I’d seen the power of the app before and have been way inspired by how the heroic Steven Johnson is using it, but the learning and experience curves always seemed just a bit steep for me, given the returns that it yielded in my too-brief usage. Still, I was quite smitten with the concept.

Flash forward a year and a half. I’ve now had DT Pro v. 1.1.1 in battlefield action for the last few weeks, and have been dutifully feeding it anything I find that seems tangentially interesting or useful; a few custom Quicksilver triggers mean one-click, no-look addition of any data type, from web pages to text selections to photos, full PDFs, and movie files. Thus far, this includes stuff like:

• most of the more interesting contents of my hard drive (transparently “synced” with DT every week or so)
• all the text files in which I “live” (over 300 — also synced)
• all my Safari bookmarks (over 3000)
• all my del.icio.us links (also over 3000)
• full text of all my 43 Folders posts (over 400)
• full PDFs and excerpts from a ton of books, manuals, and slide shows
• RTFDs or full web archives of over 100 interesting wikipedia pages (this is definitely the fastest growing sector)
• any interesting quotes, quips, snarks, canards, nuggets, scraps, emails, web pages, or random ephemera that cross my transom

My focus over this time has been strictly on capture, rather than trying to make anything particularly useful of it all just yet. But I’ve recently started grouping and classifying occasional clusters of content using the app’s killer feature: really smart AI that finds associations between items based on a concordance of common words and similar previous relationships you’ve established.

So, I have the start of a potential post underway that will re-introduce DT in more detail (which I’ve been building right in DT, natch), but I was moved today to share the insane usefulness of DEVONthink’s “Smart Groups.”

After weeks of mumbling to myself about the need for tagging in DT (and trying a few interesting and useful experiments), I finally did a little research (a/k/a “read the manual”) and realized that several of the (metric assload of) AppleScripts available for use in the current edition can create a simplified version of this for me. The scripts are drop-dead easy to use. You just create a new, blank group, title it the word or words you’re tracking, and then link to the appropriate AppleScript from the “script” field of the group’s “Get Info” window (see image to the right). This does a serviceable job of creating and automagically updating collections with all matching docs each time you click on it.

So, in the example illustrated at the top, my “knowledgeworkers” group creates a folder full of “replicants” (think: OS X aliases, not Roy Batty) for any docs that contain the string “knowledgeworkers.” The smashed-together names, as you might gather, shows you how you can fake tagging by adding keywords to the comments field of any documents; scripts in this thread make it super-easy to “tag” multiple documents at once.

The beauty part is you don’t have to use these ersatz tags at all; if your needs are modest and you just want to find all the items that contain both “learning” and “amygdala,” for example, it’s a breeze. The smartypants concordance eats stuff like that for lunch and excretes constantly updated results right into your designated Smart Groups buckets. Nice and easy.

Like I say, I’ll probably have more to say on DEVONthink soon, but if you find yourself dealing with very large amounts of information that need to be stored, synthesized, and re-combined, there’s a lot to love in DEVONthink. It’s had a subtle effect on the ways that I collect information, but I have a feeling its impact on what I do with it might be pretty profound.