I've been using DEVONthink Office Pro for a few months now to organize and mine the massive amounts of information—PDFs, presentations, HTML archives, article drafts, images, notes, emails etc.—that I've accumulated and generated over almost a decade (yikes!) of writing for Ars. I learned early on that the program doesn't do that much until you just dump everything into one database and start to make use of the search and classification features that it provides. So now that all my Ars stuff is in one huge and still growing database, I can go back and find, for instance, the links and documents I used in doing background work on a particular news post, or what presentations I have that relate to a particular company, technology, or topic, and so on. In short, I now spend less time using Google to try to dig up stuff that I've already seen but just can't locate and more time doing actual research and writing.
As much as I love DEVONthink, it's not without its flaws. The interface needs an overhaul to make it easier for me to do more of my own organizing and filing; for example, tags, smart folders, more views, and better metadata manipulation in general would all be nice. My recent flirtations with EagleFiler brought home for me some of the shortcomings in the present incarnation of DT, and it got me wondering where DEVONtechnologies is taking the product. So, I fired off an email to the company requesting an interview, and DEVONtechnologies' president, Eric Boehnisch-Volkmann, was kind enough to indulge my questions.
Eric not only told me where DT is headed, but he also told me some things about the product that I didn't know. There are a few features tucked away in there that I wasn't aware of, features that will help tide me over until the next major release.
JS: Before I dive into details of some specific features I'd like to ask about, I'd like to start with having you articulate your big-picture vision for DEVONthink development. What general directions do you plan to move the product in?
EBV: For us, DEVONthink is a data hub, a central repository for all the documents and items one works with on a daily basis. Contrary to a simple file manager like the Finder, it knows about the documents and helps the user to work with them, e.g., by providing assistance with filing new documents, organizing them, and finding the one document that the user needs right at this moment.
In the future, DEVONthink will be easier to use and come with more functionality that enables the user to use it in a workgroup. Except for being a personal database it can then be used by multiple people to collect, share, and publish data.
JS: The most useful things about DT for me are definitely the search and sorting features. These set DT apart from similar applications. Can you tell us any more about how this "AI" component works? Is it pretty much where you want it to be, or are you still putting a lot of effort into refining it?
EBV: The AI is not a component but the basic structure of the database itself. All information that is stored in DEVONthink is therefore by default analyzed and classified in the very second is added to the database structure. We are constantly working to make the AI better, to increase its effectiveness, to reduce the memory footprint, and to improve the database structure. We are also developing new products based on our technology from which improvements will flow into DEVONthink.
JS: My biggest gripe about DT is the limited number of file formats that the dB supports. I use OmniOutliner and Mellel, and I really wish that I could at least drop those files into the dB for storage, just to have all my project files in one place. Any plans to open the dB up to different file formats in a future revision? [Apparently, I was missing an option in the Preferences panel.]
EBV: DEVONthink can store any file even if it doesn't know about the file format. The user can add the file and open it with a double-click in the default application for its type. We are also planning to add support for more file formats, e.g., OpenDocument or other XML-based file formats for which XSLT templates are available that allow us to read and render them.
JS: My second biggest gripe is the way that DT handles email. I've been playing around with EagleFiler recently, and it's great to be able to search on fields in the email header (from, to, subj., etc.). Any plans for better email header awareness in DT?
EBV: Yes, but this requires substantial changes to the high-level database structure. We will make these changes with DEVONthink 2.0, for which no release date has been fixed yet.
JS: On a related note, any plans for MailTags integration?
EBV: MailTags is already supported, tags and project information is copied into the Comments field in DEVONthink.
JS: What's your vision for the DT interface? Will it get more of an overhaul in the next version, and if so, can you tell us anything about it? Or will there be incremental improvements to it? Maybe a tagging feature, or other kinds of metadata-based improvements?
EBV: The interface will be an overhaul. It will be easier to use, look nicer, and hide more functionality that is not for Mr. Everybody. Tagging and customizable meta-data are on our list as well as many other new features that make the DEVONthink workflow smoother and elegant.
JS: Two words: live queries. One day, perhaps?
EBV: Yes, that will be possible eventually.
JS: Will DT be sticking with its .dtbase format?
EBV: Yes, because the .dtbase format is the basis of all the AI. But, DEVONthink 2.0 will store all documents also in their original format inside the database package and use the AI structure only as an index. This makes it possible to search the database also with Spotlight and easier integrate third-party file formats.
JS: Can you tell us what new features of Leopard, if any, you plan to exploit in DT?
EBV:That is hard to tell. We will definitely support full 64 bit as well as multiple processors and cores, but we may also exploit some of the user interface improvements as well as Core Animation.
Thanks again to Eric for replying to my questions. I'm sure he'll be monitoring the discussion thread attached to this post, so if you have feedback/questions of your own, drop in and ask them.
reader comments