Seriously trying the DT way, but not quite there

Some background: I’m trying the demo of DEVONthink. All of my current text and bookmark content is in NoteTaker, and my more graphical diagrams in Curio. I’ve made (pretty much) a duplicate of my stored content and bookmarks into DT. This of course has meant the investment of a lot of hours of organizing, reformatting, etc. Everything seems to be there. Searching is fine, etc. Just to make clear that I am really giving this a go, and not creating my review based on a light examination.

I’m a professional technologist. My content is everything from articles, to code-snips, to server configurations, passwords, software registrations… It is obviously important to be able to review and maintain this stuff in a single application. NoteTaker does a pretty good job, but has a clumsy search implimentation, and a limited concept of the shape of data. I’m feeling the need to find something better - which doesn’t seem to exist.

Migrating to a new tool for this task is a big commitment. Not in terms of money, but in terms of workflow. It is not a decision I take lightly at all, because I am entrusting the care of my company’s most valuable data to it as well. It leaves me in a weird place, because I can’t easily go forward on my data creation until I know where it will really be “living”. I need to create a workflow around some new tool which I can expect to use for the forseeable future. To that end, I’ve made the rounds again. Looking at all the available tools and alternatives.

In many ways, DT seems to improve on the alternatives. It is versatile, fast, scriptable, and has excellent search functionality. So what is the problem then?

The problem is that DT seems to make versatile, fast, scriptable, and searchable chaos out of chaos. Still chaos, but in different ways. The tree system is fine, but double-clicking items opens things in all different kinds of windows. Navigating in DT seems like it is using intergalactic stargates of something. “Like wha? Why am I in icon view all of a sudden? Where am I in my data? I don’t recognize this place.”

The treeview… where depending on which version of it you have activated, determines how it functions and what it displays. Looking at a folder… er, group for instance, sometimes I can see its contained documents, other times it appears to be just an empty folder. If I turn on the item count functionality, this is recursive too… so a group with a single item in it will have 604 next to it because the item is a group which contains other groups. It seems odd to have something with 604 next to it open to reveal only one thing… which says 603 next to it.

Sheets (a common technique used by several competitors) seem like a great idea, but in DT they are somehow funky. The single-font and no hot-text links. No additional functions for drag and drop, no easy copy buttons to grab passwords or whatever from a sheets field.

Lack of connectivity to other OSX built-in apps. Why the need for all the weird workarounds to work with addressbook, iCal, and Mail? Contacts and reminders seem like they are important parts of data linkages to me. Other applications seem to be able to do this a LOT more transparently.

Why can’t I create a link to a folder without linking the entire (recursive) contents of the folder?

Why doesn’t DT autorecognize URLs and MailTo’s in text data?

Now, I realize this is v1. Because of this, I’m on the fence about DT. It does a lot of things right, but then at other times, DT itself becomes the obstacle. Particularly when it does things like switching the view mode and not even showing you where in the view the item you are viewing is located. Or when you drag in an iCal event, and it just vanishes… and you have to think about why, and how to actually accomplish something that will work similarly to what you were trying to do in a mac-natural way in the first place.

I can pretty much get past these issues, by creating a special process or automator workflow, applescript, whatever… but I’d like to have a better sense of why. What is the roadmap for this tool? Where is it going, and is that where I want my data to be? Will there be more and more semi-complete features and proprietary implimentations? I’ve seen several mentions on this forum about v1.1, and as a person on the fence, who needs to make a decision, I’ve very interested in what parts of this application are considered to be complete, and what parts will be getting further attention. The technology is great. The interface techniques, Mac integration, and user experience can be improved a lot.

Please don’t get me wrong. The product is quite nice, and probably meets the needs of many. Even for me, it is an obvious front-runner. But, I haven’t found a level of comfort with it yet, and I’m not sure that additional use of v1 will help that. At this point I have a neutral and almost adversarial relationship with my data management. :cry: I’d like to get past that, and into something more like a partnership, since these are the kinds of tools that you basically move into and live with for a long long time.

Thoughts, comments and roadmaps appreciated. :slight_smile:

Hi, Tryll:

Welcome to the discussions.

Yes, DT Pro is in version 1 – but with four years of intensive development.

It allows a great range of user options for views and for data import. That can be confusing, and version 1.1 will simplify the import options.

Many of the features – such as numbering of group contents – are in response to user requests. Example: How many references do I have in this topical group? In your example case, “604” is much more informative than “1”.

Perhaps it may help to think of DT Pro’s development directions as giving weight to two functions: document management and information analysis.

Document management: DT Pro does a good job of this, as it can handle very large collections of documents in a variety of common file types. I’ve got tens of thousands of references in my databases and still get very fast searches. On my PowerMac I generally get search speeds of 50 milliseconds or less for exact searches of phrases. My databases are very stable, although I always recommend backups. I’ve normally got 5 or 6 windows open, such as the top-most view, several individual groups such as bookmark and project groups and several document windows. So what might at first appear chaos becomes transparency – access to the information in various levels of aggregation, then to particular documents.

Information analysis: This is what makes DT Pro stand out from other applications and makes it a good research assistant. DT Pro not only knows where all the documents are; it also ‘knows’ all the words in each document, and all the words in the entire database and can recognize patterns – contextual relationships – of word use. While viewing a document, press the “See Also” button, and DT Pro will suggest related documents. Or select a section of text and Control-click to access the “See Selected Text” option, which behaves much like “See Also” for a segment of a document (this operation works well in the beta of DT Pro 1.1 I’m using, but not well in version 1.0.2). Or Option-click on a selected word, which displays a list of documents containing the term. I always caution about “See Also” that the user has to understand the topic being explored, but features such as this help one get at the information content of documents, which is to say, such features help one think about the material. I often discover relationships that I wouldn’t have thought of, otherwise.

So DT Pro has a lot of development effort devoted to functions that simply are not present in most applications.

DT Pro isn’t perfect, and this very active user forum is a good place to raise issues and make suggestions.

Dive in, work your database and see what you think.

Thanks for the encouragement. The items I’ve clipped above remain my concerns. I’m also re-evaulating OmniOutliner Pro (it would be an upgrade for me), since I’ve decided that NoteTaker is not handling my data well. OOP has the integration advantage. It is no where near as powerful “under the hood”, but it has a more elegant user experience, IMO. No, it isn’t the same product as DT at all, but in terms of archiving and searching through notes, they do a similar job (at least from the user perspective.) It isn’t a browser either, something I’m not sure I need in the first place. I tried DEVONagent, but pretty quickly realized it wasn’t designed for me, someone who doesn’t have much problem finding what I’m looking for on the net. OOP also doesn’t try to offer a record-style catalog of sheets. Something I’d like, if it was well designed. Currently DT’s offering in that area is pretty primitive.

The way I’m seeing it, DT has a more powerful engine and more overall functionality. OOP has a more refined user experience and better integration with standard OSX applications. I like both tools actually. However to me, they are mutually exclusive since I need them both for the same job. I’m thinking though that neither tool is actually designed with users like myself in mind, but DT more so for “infoworkers”, and OOP “just for outlines”. For me, I guess my needs land somewhere in between, and while DT has more in the plus category, it also has more in the minuses.

I’ve spent a few days solid now trying to design myself into a workflow with DT at the center, and have been unsuccessful. I don’t think this reflects badly on the tool, just that it isn’t meeting my personal needs or expectations. I constantly try to do things with it which it doesn’t support. Dragging in all sorts of items, for instance (drags that both OOP and NoteTaker handle with ease), in hopes that it will work and I can move on to the next task. Unfortunately, it generally doesn’t, so I’d have to contrive a “process” for how to keep my data synchronized, integrated, etc. I really don’t want to do that.

Still on the fence, and still experimenting… but my time dedicated to the task is finite, and I’m going to have to make some kind of decision for my existing data and upcoming projects.

Hi Tyrll
Can you specify your drag and drop problems? In my experience drag and drop to the dock icon (instead of the application window) always works if it is a file. For the clippings same method works as well but I prefer the services keyboard shortcut.

For your concerns about the navigation, the solution I use is to use the three pane view and never and ever double-click. So that I always see where I am, what are the contents of the folder and the selected document. If I need to more space for the document I go to the “Full Screen” mode.

Also I agree with Bill that showing file counting recursively is more informative.

I hope these help you a little bit, even though I feel that I didn’t understand your needs very well.


DT Pro uses standard Cocoa text routines. Links can be recognized in rich text but not in plain text, unless one uses IceCoffee.

I won’t try to influence your choice of working environment, as you are the best judge of that. If your collections of material are relatively small, OO should work for you. If your collections grow into the hundreds of megabytes, DT Pro works well, especially for researching the material. It depends on what you do.

Same with Web browsing. DT Pro’s browser has excellent routines for Web captures, as does DEVONagent. If I want to look up the web site of a camera company, I’ll use Google. But if I want to pull down all the reviews of a digital camera, I’ll use DEVONagent, as it actually captures the reviews to my hard drive. Not long ago, I used DEVONagent to pull down over 10,000 pages for a project and then transferred the material to a DT Pro database. The combination let me use the information in ways much more powerful than anything else that I know of.

LOL :slight_smile:

I think you hit upon the key usability weakness in DT.
It’s a great application with TONS of potential to be the killer app for OS X, IMO.
The GUI and navigation do seem chaotic and unnecessarily complicated however.

Perhaps this is inherent in the versatility of DT but I cannot help but feel that there is plenty of room for improvement in that area.

I have lots of confidence in the Devs as well. :slight_smile:

Yeah, seems that more than a few find themselves confused by the way navigation is handled in this tool.

As for me, a demo user, I’m going to take a pass on DT for now. I have at the very least a temporary home for all of my data in Curio (another excellent product) and the native file system/spotlight (fast and flexible), and I plan to watch and see what happens with DT over time.

Yeah, that got me at first until I learned what physicistjedi said, “never double-click on anything”. This usually happens while browsing through the group hierarchy … the switch from tree view to icon view is disconcerting . Took awhile before I realized the ‘home’ button could sort of be used as a ‘back’ button in this instance, though a true back button (‘previous view’) would be better.

I’ve noticed there are a few things about DT that take getting used to, which usually crop up whenever I am showing someone the wonders of this new tool. All to be expected of course, every user comes to the app with different expectations and different habits.

Perhaps a ‘UI annoyances’ thread should be started. I know I’ve got a few, like the way groups don’t automatically expand when groups are created inside of them (gotta arrow -> in to name the new group).

As milhouse says, a great app – in fact it has almost single-handedly brought the first macs into our office – but the UI needs some usability feedback so that new users feel unquestionably good about using it.

We can’t state it enough: we love feedback and this thread is very useful to us.

The application is very powerful and it takes some efforts to come up with a user-friendly way to present all this power. We’re still working on it. For example, the upcoming release will make importing file data much more transparent.

So let us know how our user-interface surprises you in unfortunate ways. We are listening!

I really like the option to double click on a group to open up a new window. Makes total sense for me in implementing a GTD like workflow in one database – double click on “Next Actions,” double click on “Reference,” and I have two windows to crossreference. I wouldn’t want it any other way.

And I don’t get lost, thanks to the implementation of command-R “reveal.” This ensures I never really feel lost.

I also find the Groups floating palette handy, though I wish it had more minimizing options a la Drag Thing…

One suggestion I have for retaining a good sense of where in the hierarchy a document lives is a “Path Finder” style breadcrumb trail at the top of either a document window or a group window.

Example of Path Finder’s breadcrumbs is visible here – just below the tabs in the interface, a string that reads “alec | Utilisateurs | neilio | Images | Stock Photography.” If you haven’t used Path Finder, know that each spot in the breadcrumb trail is clickable and draggable.

Hope that makes sense – it’s very powerful.

A simple fix for the double-clicking issue might be to enable the “Open groups in new windows” preference by default. Then double-clicking will open a new window. What do you think?

I think that’s a great idea. Having it open in-place, replacing the three-pane view with a single-pane view, is what’s confusing. In fact, let me enable that right now :slight_smile:

The breadcrumb idea is worth considering, possibly as an extra line down by the classify buttons… though real estate is already scarce down there. As it usually goes in programming, “sounds great! … make it an option.”

I think I’ll wait and see what 1.1 brings to the table. :wink:

I wonder if Bill (or anyone) can explain this a bit. How can one not have both documents and subgroups living inside groups? (Or maybe my question is, why should one not?) My data is all organized this way: every group has some important top-level documents and some subgroups for the rest.

I share the original poster’s concern about the feeling of the GUI (much as I love the program), and I wonder if this has to do with why.

For example, I do feel I have to use the three-panes view most of the time. The horizontal-split view shows subgroups but not documents (via disclosure triangles) in the top pane, and I virtually never want to see documents as batches of icons. The vertical-split view does show both subgroups and documents (isn’t this inconsistent?). In both of those (the split views), it seems weird that the second pane shows group contents only in icon mode, rather than list mode.

Or is there a preference option I’m missing?

The theory of hierarchical organization is that one should not mix documents and subgroups inside an enclosing group. To do that is an indication that either the organizational structure isn’t complete, or that some documents haven’t been filed properly – and it will to some degree confuse the AI Classification feature.

As for me, I break that rule all the time, and I never have the time or inclination to try to clean up my organization. :slight_smile:

I chuckled when I read about the confusion because icons are displayed in the Vertical Split view when one selects a group. That’s because, although I use the Vertical Split view in my work, I don’t see that behavior, and had almost forgotten that it can happen. Here’s why:

  • Although I keep the top-most view open, I rarely work in it. Instead, I’ll double-click on topic or project groups to open them in their individual windows and work in those environments. (I’ve set Preferences to open groups in new windows when they are double-clicked.)

  • In practice, I don’t select a group. Instead, I click on the disclosure triangle to inspect the contents.

So, for me, the Vertical Split view provides an entirely consistent work environment. I prefer reading or writing in that view, because I like to have the maximum screen height for my documents. And I’ve set windows so that the Info panel is always “attached” to the right of the text pane, and is resized to stretch from top to bottom of the screen. So I can easily enter notes and comments in the Info panel’s Comment field (especially convenient for PDF, images, and “unknown” file types, which can’t be edited in the text pane). Another advantage is that, to edit a document under it’s native application, I just click on the Path icon in the Info panel – or launch an HTML file by clicking on the URL icon. And of course I’ve told Preferences to reopen my working set of windows next time the database is opened. So my familiar set of windows reopens, sized and configured just as I’ve described.

And if I’m reading or writing a long document, I double-click on it to open it in its own window (with the automatically “attached” Info panel). That way, I can rummage about looking at other stuff in my database without losing my scrolling position.

As for me, because of my preference for the text pane of the Vertical Split view, I find the Horizontal Split and Vertical Split views strange and confining. Different strokes for different folks. DT Pro provides options among which you can develop a comfortable and consistent working environment.

And yes, the Vertical Split view does allow one to “mix” groups and individual documents. I’ll sometimes put my “To-Do” and reminder documents there for easy access in my current project group. If I’ve got a big group that’s sloppy, with lots of mixed groups and documents, I’ll sometimes temporarily switch to Three-Panes view to find the group I’m looking for, open it, then switch back to the Vertical Split view (hey, that’s where the icons pop up, to highlight my chosen group).

When DT Pro 1.1 is released, I’ll try to remember to post a really useful tip about selecting a group or subgroup to receive new imports. Those icons provide useful feedback.

Thank you, Bill – that’s all good. It helps me see what my complaint is.

In your style, working in the vertical split view, isn’t it still awkward and inconsistent that if you select a subgroup in the left pane, you’re forced to use the Icon view in the main pane? I would strongly prefer a List view almost all the time. Logically, I think I should have a choice.

Also, when I do have documents mixed with subgroups (which is pretty much always), I feel the subgroups should be sorted at the top of the list, not in alphabetical order among the documents.



No, I’m not forced to use the icon view in the main pane when I select a subgroup in the left pane. Why? Because I don’t select a subgroup in the left pane. I either double-click on it, or click on the “reveal” triangle. Because I only perform actions that are meaningful and useful to me, the workflow is entirely consistent and efficient.

So my workflow isn’t awkward or inconsistent. Using it, the icons never appear. (But, as I hinted, those icons do come in handy for another purpose involving sending new content to a selected group.)

Because DT Pro remembers your window sizes and locations (and you can adjust those to your satisfaction), it’s simple to “attach” the Info panel to the right of the Vertical Split view window, and also to document and Search windows.

See a picture, “Modified Vertical Split.jpg” at

You want to look at the left pane that way? OK – you’ve got it. Just select View > Sort > Kind. That does it. Want to look at the left pane in a different way? You’ve got lots of options.

I love the flexibility DT Pro provides this way. Much of the time, I’ve got the sort set for Modification Date. But sometimes I want it alphabetical. Or I may sort by Word Count. Or with no imposed sort, letting me drag items into just the order I want.

Want still more of that kind of flexibility? Use View > Columns > … This lets you add one or more additional columns in the left pane. I often do that, since it allows quick and easy switching of sorts. For your case of a view with lots of mixed documents and groups, just add a column for Kind. Click on its header and you will instantly segregate the groups from the documents. Reveal the contents of the group you want and switch back to an alphabetical, modification date or other sort of your choice.

Christian has put a lot of thought into the GUI. It’s not chaotic. It’s not inconsistent. The GUI options are deep and rich, allowing you to customize DT Pro’s working environment as you like. :slight_smile:

Instead of forcing everyone into a single, fixed environment, you’ve got GUI “tools” that you can assemble and use for your personal taste. Play with those tools and you may find a work environment that you love, and that’s different from mine. This is another of the features that made me a DT/DT Pro user.

Excellent! Thank you. I’d missed this.

You’re saying there’s no interface problem because you simply avoid that action (select a subgroup in the left pane). But surely selecting a subgroup is a normal action. When I do, I expect (or desire) to see a list of the subgroup’s contents in the main pane. (Yes, I know the disclosure triangle is an alternative.) It feels as though that should be an option.

Thanks again.

I have to respectfully disagree. Simply because you have adapted a work style which is supportive of DT’s peculiarities, does not mean that DT’s UI functions intuitively. I don’t question the program’s power, but the UI does not work as well for everyone, particularly new and demo users, as it apparently does for you. The initial behaviors of the application, and then even with ongoing (trial) use, left me feeling that although my data was well tended, it wasn’t any easier to work with than other (equally flawed) systems which I already own. Yes, I could make myself work with it if I had to, but I’d rather use something which at least met me halfway. DT currently doesn’t, for all the reasons I mentioned at the start of the thread, and probably more.

This is not to say that I don’t respect the amount of work and thought which has been put into the program, only that it still doesn’t meet my own expectations or compliment my work flow. For instance, in addition to UI issues, DT does not integrate well with the OS, doesn’t support iCal, Mail or Address Book well. While this could be blamed on Apple, the fact is, virtually every competitive product does these things, so how much of an obstacle can it really be? In fact, I am in the process today of integrating one of my own applications with Mail, so… ?

Obviously, I see a lot of potential in DT (that’s why I’m still here), and hope that one day it becomes something I want to use, but I’m currently not willing to give up on the concept of something better, with a more solid UI, and better integration with the rest of my tools.


I tend to agree as well. It seems that the program’s UI is not designed so much for the way the users tend to work (whether or not it is the way the program is intended to be used or not).
It is very powerful but, as much as I do love the program, it isn’t quite user friendly.

Tinderbox is another very powerful application and the usability borders on bewildering especially for novel users. Perhaps it is the nature of such programs?

It’s almost as if DT needs several different UI’s based on the current usage scenario.