What does the future hold for DEVONthink?

I would want to be able to add custom property/value pairs to a DEVONthink record (document or group). I.e., “custom metadata”. And then have these data available to views, Search, and the scripting dictionary.

I want more support for LibreOffice OpenOffice formats. I’d like to drag my LO files into DT and have them visible there without having to have LO on the machine. Most useful for viewing documents on my iPad.

I agree on the more macro/scripting ability similar toHazel.

Also agree on the separator pages during OCR, that would be really useful.

I disagree on changing the user interface too much. I personally do not like the “upgraded” interfaces that iOS7 brought and the changes that companies like Omnifocus made to their products as a result. Just because it’s an old interface doesn’t mean it’s not still a good one. I hate companies that make changes just to make changes and especially when the changes affect visual appearance without any benefits. Style is unimportant to me, functionality is.

I have no experience with LibreOffice, however if you can view LibreOffice documents in DEVONthink when LibreOffice is installed, but cannot view them when LibreOffice is not installed, that would suggest that LibreOffice has a QuickLook .qlgenerator file inside its application bundle. You can copy and paste that .qlgenerator file into your ~/User/Library/QuickLook folder and you will not longer need to have LibreOffice installed.

On metadata

Definitely want to see custom property/value pairs. The ability to add custom meta data to a record greatly increases its usefulness in many situations (and for what I do is pretty much essential). Presently I have to do this with lots of cumbersome applescript parsing of spotlight comments

On changes to the UI

I understand the desire to freshen up the interface and make it more accessible to new users but honestly I wouldn’t know where to begin. Much of the virtue of the present UI is that is stays out the way and lets you focus on the documents which is honestly what you want. More ‘chrome’ would be a distraction to getting things done.


Overall the documentation is superb but I would love a section on the website dedicated to the best applescript from the forums - perhaps a wiki format to which we could contribute. Something like http://en.wikibooks.org/wiki/AppleScript_Programming but focused on Devonthink applescript. I have spent many hours searching the years of accumulated wisdom and studying the many gems which lurk lost (thanks korm over and over again!). The incredible customisation of Devonthink is its greatest strength and promoting the flexibility should definitely be more prominent.

There seems to be renewed focus in Apple on applescript with AppleScriptObjC. I have just begun to study it in detail but it seems to be a direction Devonthink’s applescript should be moving.

It was after seeing someone demonstrate Devonthink I bought a mac so my continued thanks to the whole team.


The ability to search and manage epub and mobi like we can pdf.

I know mobi is a long shot, but epub is pretty open.

I wouldn’t either, but then I don’t have to. The present UI is too crowded, and for all the choices of layout, if you have groups or documents with names longer that 8 or 10 characters the only way to view them comfortably is in List view. I don’t understand why command-i won’t open the Info pane like it does in the Finder, iTunes, iPhoto etc. The ability to set margins on notes and text files would be appreciated and if… oh if…

you could roll a proper citation manager in there it would be wonderful.

And yes, I’m aware of the “integration” with Bookends but yet to find a use for it.

Nope, not viewable at all within DT. Have to open into the Libre Office program directly so no quicklook thing in there.

Providing a Quick Look plugin to allow Spotight indexing and Quick Look display would be a job for the developers of that software suite, not for the developers of DEVONthink.


Support for epub (or at least for non-DRM epub files) would totally rock! Note that epub is remarkably close to html in oh so many ways. Please, oh please!!!

A proper citation manager would be cool. One possible approach would be to build on top of latex .bib files. These have the charm of being (a) well-specified, and (b) importable to and exportable from nearly every other citation manager around.


…no longer relevant…

better DTGO, better sync and modern UI.

I would like to see an extra sync capability: Setting (per database) a One-Way Sync.

For example - we keep a Master database, and then our remote users can sync (download new data), but should never be able to overwrite (upload) back to the Master. Sometimes they do by mistake, and then we have to clean up…

Personally I’m more than happy with the UI. Considering the capability and underlying complexity, the UI does a pretty good job making it accessible.

I think DT could do with a quality tutorial linked to a demo database for showcasing the best capabilities to attract new users. Beginners are usually overwhelmed - unable to see the wood for the trees.

DT is incredible software - but is almost a well kept secret…

I too am generally satisfied with the UI. I would tweak a couple of things:

  1. View > As Tags shows all the tags in all databases, and when I’m looking at a single database I don’t need to see the non-applicable tags. OTOH, it would be great if View > As Tags has a mode switch so that I could browse by tag across all open databases – in other words, in one mode I just browse by tags in this database, and in another mode I look at all databases
  2. I would like to be able to dock Show Info – it’s really an Inspector panel, and could have give access to a lot more features than it currently does, such as: a tab for text formatting, a tab for annotation, a tab for Properties (Tools > Show Properties), a scratchpad tab for quick note taking

FWIW, good suggestions/requests these - would welcome these as-well. If the Show-Info was docked – interesting possibility could be similar to quick-view in Finder – how it ‘updates’ as you switch files in the background.

I’ve been using DEVONthink for over 7 years and regularly come to the forum to keep tabs on the conversation but up until now I haven’t actually posted anything here. This thread however sparked me into action as I think that DT3 and DT2go are critical upgrades for DEVONtechnologies as the computing world is finally leveraging the promise of ‘the Cloud’, Evernote has filled a gap for the broader consumer end of the market and much as I love the strengths of the DT suite of tools the underlying software design was created at a time when cloud computing had a long way to go.

Before responding to how I’d like to see DEVONthink improve in the future I think it would be useful to give those thoughts some context with regard to how I use it at the moment.

I’m a design strategist/UX consultant by trade and I maintain a streamlined DT database that’s functions as follows:

  • An index of all my client project documentation, kept in Markdown format (so that it can easily be republished (with appropriate styling applied) to web, intranet, PDF, and print formats from the same Markdown source material).

  • My web research archive, imported from Pinboard. I capture the HTML content from the .webloc bookmarks so that I can take advantage of DT’s advanced search functions on the content without the need to add unnecessarily heavy media files to my database

  • I manage and merge my tag classifications through DT. The tags are imported from Pinboard and I regularly export the Pinboard folder to my Mavericks home/documents folder so that Mavericks registers the tags

  • I use Ammonite to get a unified view of all my tagged documents whether they live in DT, DropBox or an OmniPresence folder (for various Omni Group doc formats)

  • I integrate an index of a very large PDF library. The PDF’s are annotated in Skim and I use a Rob Trew applescript to import those annotation notes to DT (it also imports an alias link to the PDF in question which means the annotation notes can directly open the PDF in question on the relevant page.

  • I have a ‘live project folders’ which mainly contain replicates of existing DT data that’s relevant to that particular project and alias links to project documentation in a variety of formats (Omni, Adobe, MS, Apple etc).

My use of DT is totally based on an an indexing workflow. I never import documents, PDF’s, media files or such like. I’m not suggesting for a second that an indexing workflow is better than an import one (I’ve witnessed the debate that ensues here on the DT forum to know better than that) but it’s the workflow that provides the best results for me.

DropBox is the glue of my productivity workflows but I also maintain a media server (to host all the rich media referenced in my Markdown docs).

The most important thing to me is that no piece of software or web service I use attempts to impose ‘monolithic’ rules to the manner in which I work with my data/documents. Mavericks tags makes it really easy for me to associate documents to a specific project without having to make use of a traditional nested folder approach. This is important to me as I work across desktop and mobile laptop each day and it’s easiest for me to keep core project management tasks in the finder and filed by document type rather than project.

IOS devices are becoming increasingly important to me for productivity purposes too. For instance, all my meeting notes are created in OmniOutliner on my iPad mini and end up as Markdown formatted notes in Day One which I use as a ‘day book’). I’d love to include DT2go in my mobile workflows but it adds no value at all as it stands. Over the last couple of years IOS devices have become productivity workhorses because of protocols like x-callback and applications like Editorial that are able to leverage x-callback. I really hope the DEVONtechnologies team take a lead from this collaborative approach to working with files on IOS and ensure that DT2go functions as a hub application not a ‘monolithic’ one.

On to the things I’d like to see improved. There’s been a lot of great stuff posted already so I’ll add comments to the points that have already been raised.

The main reason I use DEVONthink rather than Evernote (or the full text search capabilities of Pinboard) is that DT offers far more sophisticated search options. I’m not sure it’s so much a matter of improving those search options (although they could benefit from some of the tweaks @Cassady mentions) but the user experience could definitely be improved. I personally hate having to open the advanced search window for such a critical function as search by tag. Whilst Ammonite has filled a gap here it’s still a filtering utility not a search one. The positioning of the advanced button on the search window is a particular UX bug bear as it’s so critical to sophisticated search queries. I’d like it to be as close to the text entry field as possible.

Whole heartedly agree with this and look forward to what Christian has cooking (from his not so cryptic response).

Whilst i’m in agreement that more chrome isn’t the answer and the interface needs to keep out of the way as much as possible the UX is far from ideal. Omni Group may be getting flack from their hardened ‘fan-boy’ user base of old with the UI refreshes of the last few years, but the vast majority of users have voiced positive feedback with regard to the usability. I’m no fan of MS but the ribbon was a fantastic UI strategy that manages to conceal complexity and reveal deep functionality in an intuitive manner. DT already provides so much power, I think the challenge is to reveal that power at point of need without getting in the way.

Full agreement here. ePub is increasingly being used to distribute project documentation (because it’s effectively a packaged HTML file) so it would be great to add ePub files to my index.

Overall my hopes for DT going forward is that it it becomes:

  • More cloud friendly,
  • The indexing side of the feature set is given equal billing
  • The UX is improved in an unobtrusive manner

My final wish is quite a radical departure and it takes us into Evernote territory. I’d like to see a hosted version of DEVONthink to be made available where you access your data through a client that can sit on multiple devices. I’m not suggesting that DEVONtechnologies provide that hosting platform but those of us that wish to host our own server version of DT can. Effectively what you already offer in DTPO but with bespoke client software for IOS & OS X rather than a web application in the browser.

Apologies for the marathon posting - I had seven years to make up for! :wink:


YES!!! :clap hands:

That’s one I keep on forgetting about, but every darn time I need to go alllllll the way down into Advanced to call up Tags etc., I find myself wondering why on earth it’s so hidden away – 1st world problems… :wink:

@jm - that’s quite a debut 8) We’re all lucky you got the 7-year itch and shared your excellent thoughts. Your posting encouraged me to print this thread to disk for permanent reference.

This is the most intriguing aspect of your post (and many others in this thread). For my own work, I am getting wary of cloud services. I have lots of data spread across Dropbox (private account), Box (business account - which I’ve come to regret), and OneDrive (also business - to replace Box and because my work requires Windows). I am evaluating options to back out of all those services and either host a private “cloud-like” server (e.g., Macmincolo) or shift to a cloud provider with greater long-term security and reliability than these three.

For years, I have grown numerous DEVONthink databases with the “import” model. But that model is now less flexible than I need. It will be a major undertaking to shift all my data into indexed mode, due to my idiosyncratic structures and reliance on replicants for workflow management, but the payback in terms of near-seamless access to core research + notes + original work across desktop, laptop and tablet platforms is attractive.

On the iPad, I moved off of DEVONthink to Go a long time ago. At first I thought I was moving away from it because of it’s lack of features. Over time, I’ve found that even if DTTG’s feature set was broader, I still wouldn’t use it because I have learned that feature-dense apps on iOS are not useful. The data are more important than the applications.

So, my emerging personal architecture is:

  1. Separate data from function: e.g., move the data out of DEVONthink’s database packages and just use DEVONthink for AI and metadata-based features, but not file management
  2. Move data into as private a cloud space as I can manage
  3. Seek and use those apps that can deal with #2 well (this is a tough one on iOS – and easy on desktop / laptop platforms)

To achieve #1, I need a few improvements to DEVONthink indexing to make it as seamless as imported data:

  1. Locate “missing” files for me; or, if that’s not possible, provide a feature for me to reconnect files to their internal database pointers – somewhat like the feature in Lightroom to tell Lightroom where missing images have moved.
  2. Keep indexed document metadata in sync with the physical file – tags, labels, comments – without requiring my intervention

So, back to @jm’s essay. I don’t think a “hosted version of DEVONthink” is what I need. I would flip that suggestion around – if the “hosted” piece is the data, then what is needed is platform versions of DEVONthink (single-user desktop/laptop, tablet (iOS or ??), or workgroup server) that can work with hosted data. I want an instance of DEVONhink on every platform I use that I can point to a single private data cloud. This takes DEVONthink out of the old-school file management and sync business, and focuses it on the features that are its real bread-and-butter. Which I think would be a great basis for future growth.

@korm, I am in agreement and seek a sensible, reliable, straightforward method to achieve that. Like you, I rely on replicants in addition to numerous x-devonthink:// URI’s for internal cross-linking between DT documents. Any tips/suggestions would be warmly received.

This is pretty much the way I work at the moment. I maintain a private server through Dreamhost that has various components (such as web server, media assets server and WebDav server for secure document storage), and a 100GB Dropbox. The combined cost of these is very reasonable (approx £100 per year). As you say this approach of separating the index data from the content itself allows one to keep a lean DT database of indexed data and enables you to use DT for it’s strengths - the AI & search functionality.

I’ve found with the latest builds of DT as long as you only edit your docs/metadata outside of DT the indexing is a lot more reliable at automatically keeping up with changes. But I agree this needs to be a better user experience. I understand the technical challenges the DT dev team have on the hands as the system is optimised for imported documents.

Apologies, I didn’t explain myself properly and probably confused the issue by mentioning Evernote!

We’re looking for exactly the same thing. As you state (with greater clarity) the hosted part is the index data (married with some local doc storage where appropriate). In many ways it’s more a version of DEVONtechnologies commercial product, DEVONsphere. But with a great client version both OS X and IOS.


This is in reply to your PM - the board software is being a tad ‘nanny state’ and won’t allow me to reply via PM seeing as I’m new round these parts! :slight_smile:

All I mean by ‘media server’ is the folder I have set up on my web server for media assets. Those assets can obviously be published by far more than the websites I have hosted at Dreamhost. A bespoke media asset server is overkill for my needs.

I mirror the media folder on Dreamhost , on a local network drive which is indexed by spotlight and I’ve been pretty good with metadata over the years so I can search my library with ease using HoudahSpot. HoudahSpot is a work of genius! A perfect partner to DEVONthink.

With regard to Droplr, I have no intention of stopping my subscription as I think it provides pretty good value for things I want to post and forget on social media sites and such like. I love it in particular for posting things like MD formatted notes, annotated screengrabs or audio files on my SoundCloud profile (it’s also great as a pearson URL shortening service). Far more rapid workflow than using Evernote for that kind of stuff.


This service intrigues me too, especially in this post NSA revelations world.

wired.co.uk/news/archive/201 … rom-google

If I can manage to bill the deployment to one of my clients I might just give it a go! :wink: