Difference between DT and Tinderbox?

DEVONthink Scenarios
41

smolk, your post interested me very much, as I own both DT Pro and Tinderbox, and use DT Pro regularly. Tinderbox languishes largely unused, even though I know I could get a lot out of it, mainly because of time constraints: I already probably use too many tools, and I decided I had to draw the line somewhere.

What I was hoping you could expand upon is what suite of tools you use and how they come together? From your post I understand that you collect vast amounts of large documents in DT and that somewhere along the line you use Tinderbox, too, and then you write in Ulysses. Do you reproduce most of the DT material somehow in Tinderbox. Or is it for totally different material?

Just so that you know where I’m “coming from”, I use a combination of Bookends, DevonThink, and Circus Ponies Notebook when I’m reading downloaded papers. It’s a bit time-consuming, but I haven’t been able to find a way to cut down on the hassle while maintaining functionality. I use DT to store everything I wouldn’t want to lose access to, Bookends obviously to make sure I know where things have come from and that I’ll be able to generate a reference list when I need to, and Notebook as kind of interface between the two: my main notebook is organized in a subtopic-per-page fashion with short quotes pulled from papers in DT, with the short citation info pulled from Bookends in the margin. I got this idea from the Notebook writers’ forum, and it’s important for me when I’m actually writing. For example, if I’m writing about the disadvantages of discussion forums vis a vis blogs, for example, I don’t want to go on an extended exploration in DT, I just want to pull up a page with all the info I need listed, so that I can create a properly referenced paragraph without referring to anything else. I’m kind of assuming Tinderbox serves a similar role for you, but I’d love to hear details from you.

42

It’s nice to hear how you organize things; I’m also always looking for ways to improve, although at the moment I’m quite happy with my current work-flow. Which is not dissimilar to yours, I believe.

I do not often take something from DT to Tb. But I do not use awfully many snippets in Tb anyway, apart from the blocks of texts I’m working on. I may sometimes copy them from digitally available texts, but if these do not exist, or are not satisfying, I type them myself. Tb is really a collection of bits and pieces:

  1. to keep track of literature I should incorporate; this may be a snippet from DT, from BibDesk, or the Web;
  2. to remember questions, doubts, open avenues;
  3. to organize the structure of my research and my writing;
  4. to visualize connections between
    a) the objects of my study;
    b) my own studies of these objects.
    Playing around with these visualizations also reinforce my familiarity with the data.

I have one Tinderbox-file with over 200pp of text (it started as a LaTeX-file of 120pp, and it has since doubled in size). The file was structured, so I “exploded” it in Tinderbox into separate notes. That took a few seconds to do. Then I assigned prototypes with certain attributes (keywords, texts, protagonists), and agents to retrieve specific information. Say, I want all snippets using the words “oral translation”. I may not necessarily have all relevant texts under one heading, because in some texts the main focus could well be very different, but the keywords mean that I find exactly what I want. I then have a permament lsit of the texts I want to see together for future reference. And agents update themselves: add another snippet with the relevant keyword, and the agent will show it immediately.

These agents are dear to me: I use them to select information from this pile of text. It allows me to have overview and to be confident not to forget anything. (Unless I did not tag a text…) In some cases, I copy the results into a separate file to make a mindmap and further refine relationships.

What you describe as the function of Notebook for you, comes very close I guess. It really is filling a page with disconnected thoughts and reminders and quotes, all ready and available for later use. But Tb is more flexible than NoteBook. I only use NoteBook for syllabi to be published on the web; NB has a nice way of converting a notebook into a webpage. For notetaking I find the notepad metaphor imposed far too strictly on the software: navigating between pages in NB is not a pleasure as far as I’m concerned, and so I rarely use it. The index pages are not helpful either.

As soon as I really start writing, I copy bits and pieces into Ulysses. A services short-cut might have been handy, but it’s not really a problem for me. As you see, the processes are fairly separate. But you might say that Tb is the main engine of the preparatory phase, while Ulysses is the main engine during the final phase. But in practice, I use both in tandem, and heavily, and much more intensively than DT.

DT kicks in when needed to spit up or store useful material, such as PDF-articles. I still print these. A completed article also ends up in DT (although I sometimes regret that as you end up finding your own research). I don’t like it for writing purposes, and Tb allows more freedom in organizing. But DT is my treasure grove. There’s a lot in there I may never ever see again, actually :wink:

I would have considered Mori if I had not already owned DT pro, NoteBook and Tinderbox. Mori and DT could well be a winner, too.

43

I seem to recall a number of people not classifying anything in DT Pro. I have copied my folder-habit of the finder into DT: neatly arranging almost anything. Fretting over where to put certain files. But maybe DT Pro should indeed be used differently: can anyone expand on this? (Bill deVille perhaps? I believe you said something to the effect that you initially put everything in a folder “to be classified” and eventually did not classify the materials?) What is the advantage of not classifying (apart from time benefits!), and what of classifying?

44

Thanks for the detailed account of how you use these programs, which I’m sure will be useful to many people. Eastgate have finally got serious about proper help on the web (they’ve always been good with email), so now might be a good time for anybody interested to take a look at their screencasts.

I also have used Notebook for publishing syllabi, and I also publish the notebook with quotes arranged by topic to the web so that I can refer to it from home. (I have also done the same with Tinderbox and VoodooPad in the past.)

I agree that Notebook is much less flexible, and strangely lame in some ways for a modern Cocoa application in 2006. And I’ve surprised myself by choosing it partly for that reason. I figure that once I’ve done my explorations I ought to know what quotes support what ideas. At any point in my writing there would usually be only page that could be of any use. Tinderbox and DT are incredibly prone to window proliferation in my hands, while Notebook only allows you to see one page at a time and can therefore fit alongside a writing programme. I’ve done the paste-into-Ulysses thing but found that there were so many bits to paste in that it was rather frustrating. I’d rather just get writing and refer to Notebook as necessary.

For writing, I start in Mindcad Incubator, which looks somewhat mindmappy (but without many graphical frills) but actually encourages you to gradually move ideas into a strict hierarchy, to create a complete outline, then export to OPML so that I can do the writing in OmniOutliner. Each paragraph has a title as an item in the outline, but the text of the paragraph is tucked away in the inline notes.

I guess you can understand why I don’t want to add too many other programs to the list :slight_smile:

45

Version 3.5.x undeniably brings improvements but pffff this application is really overwhelming. There’s way too much stuff between me and the data. There is always a learning curve but here no, that’s too much. Tinderbox and I we have never been friends, it has always been an obstacle. In my workflow I try to keep things as simple as possible, something I couldn’t achieve with Tinderbox. I wanted to create a dynamic to-do list, organize notes about political intelligence, music, web design. And prepare my next trip (hopefully the South Pacific in 2007, can’t wait!). Things I do much more elegantly in DT Pro, TAO and Curio. I mentioned three apps here, if I added Tinderbox to the list I would spend my whole life on my Power Mac and I already spend too much time in front of my screen. You have to stop somewhere. And the three apps that I listed are not simple. What’s more, honestly, I don’t like this concept at all. I had to think about Tinderbox all the time. A piece of software is a mean, never an end. When I dowloaded it for the first time, I thought this app must be very flexible. It is. But to me too much flexibility brings ridigity, I always felt I was trapped inside Tinderbox.

46

I’ve been wrestling with information management schemes for some time now. Currently my favorite is a blend of DEVONthink with other strategies, the reasons for which I’ll get into below. But I’m still longing to find the One Environment To Rule Them All. My current thought on this, though, is that such a thing is not practical, anymore than striving for One Vehicle to go all places would be practical.

First, let me say that DT has its shortcomings. Whoever designed the UI must hate their keyboard. Of all the applications I use, none is more keyboard unloving than DT. Creating and managing outlines is so hurtful, I simply can’t do it. I have to press 4 keys at once to make a new outline node, and afterwards I can only rearrange them with my mouse? This is 2006!

That said, I love DT. It has become my information storehouse. As I realize more and more about what it’s meant to do, I see that it’s an awesome strategy for collecting and collating tons of little bits of information – exactly the kind of stuff that gets lost in the nowhere-bin of an ordinary filing system. I used to keep little one-line text files here and there on my computer, but I never ever referred to them after they were created. Now in DT, the searching and See Also pane give me a way to access all of that random data.

But DT can’t be everything. It’s not much good when you start dealing with single, gigantic files. At that point, it starts becoming a Finder replacement, but there are better ways to do that…

What I’ve evolved is a three-part scheme for managing my information. It consists of essentially three applications:

  1. At the first level of information management is the dynamic process of creation. What I need at this level is something very quick and agile; something that makes content creation a breeze, and re-organization of that content even easier. For this, I’ve not found anything to even approach OmniOutliner Pro. Whoever designed its interface and keyboard shortcuts was inspired.

  2. Just above this level is something close to it: all the little bits of information that have a static flavor. This includes ideas that have already been written now that I’ve moved onto something else; web clippings; notes I’ve jotted down and don’t want to lose, etc. For this, I’ve found nothing better than DEVONthink Pro. And believe me, I’ve tried to move away from DT (because it’s UI bugs me sometimes), but I just keep coming back to it. It’s the best “information attic” I can find.

  3. Above this level is what I think of as “bulk data”. Big, huge pieces of data that I don’t want to get rid of. This might constitute mirrors of websites, disk images, digital instruction manuals for software I own, etc. I’ve tried putting all of this into DT, but it just bogs down my database with stuff I frankly will never access. But when I want it, I want it. To manage this level of information, I use Path Finder, Spotlight and LaunchBar.

This division of my information flow maps onto what many people experience in real life: You have your sticky notes, jot pads, back-of-the-envelope scribblings; you have your news clippings, old letters, final drafts of papers you want to index; and you have your boxes of old yearbooks, junk from childhood, and other stuff you want to keep but don’t otherwise need anytime soon.

I’ve tried to make DT be all of these things. It can’t really. I know the functionality exists, but I shy from doing so for the same reason I avoid writing in pencil: because it’s cumbersome enough that it’ll cause me to avoid doing it at all.

By contrast, OmniOutliner is so delightful to use, it actually draws the words out of me when I brainstorm. In a similar way, DT makes browsing through millions of words of information effortless. And lastly, Path Finder and Spotlight make strolling through multi-gigabyte archives pleasant even. Each application has its speciality, and they all do them very well.

I’ve looked at every other system I could download, by the way. Tinderbox is, among all of them, the biggest enigma. I spent all day today reading their website, and I left feeling like I’d witnessed a huge crowd of people waving excitedly in the air about nothing at all. The website actually presented the idea that “Ret/Type/Ret/Space/Type/Ret” was a quick and easy way to enter information. Perhaps if I’d never owned a computer before and was stuck with my pencil… How about Omni’s “Ret/Type”? I searched and searched for a “killer reason” to buy TB, but in the end found it was basically a work-intensive way to do what DT does for me with less effort.

In the end, I’ve found that my motto is this:

A good information system should require an absolute minimum of front-end work, while optimizing the amount of back-end work required for a given task based on the user’s proficiencies.

If you use the right application for the wrong purpose, you’ll find yourself violating this rule over and over again, and the resulting inertia will convince your brain to seek other ways to spend its energy. But if you do successfully marry application to purpose, you’ll experience the sort of “luft” that makes you realize: yes, indeed, there are beautiful reasons why computers are around.

47

Hi, jwiegley. A well written and thoughtful piece.

Just one comment about the DT Pro UI for lists. It’s Apple’s Cocoa UI for lists, not unique to or created by DT Pro.

Recently I’ve started working with an application that has a few remnants of computers’ olden days, before the graphical user interface.

In those days, one used the keyboard a lot to do tasks other than text entry. There were lots of keyboard commands, which could be strung together as, effectively, little applications to accomplish a task.

When the Mac was introduced it generated excitement because one could work with a graphical interface, such as dragging a list item to another location rather than typing a bunch of commands to accomplish the same thing. :slight_smile:

The application I mentioned that has some “antique” elements is Papyrus 12. I’ve absolutely fallen in love with Papyrus because it can generate editable PDF documents, and is also a quite powerful office suite. And Papyrus has good spreadsheet features including the ability to place elements of a spreadsheet table within a document that point to data in other documents. The “olden days” feeling, though, is aroused when one has to describe the location of such data. Instead of a graphical user interface in which one points at an external document, and then the cells to be incorporated, the user has to type commands like this: “E:\DATA\GET_THIS.PAP#MYLABEL]”. Followed, of course, by still other text strings to accomplish the purpose.

When I was doing a fair amount of programming the only API was the ability to specify the x,y coordinates of pixels on the screen. Everything had to be completely spelled out in code for screen views and printed reports. Even those somewhat antique requirements for specifying and capturing external data that are present in Papyrus were a great advance over my programming days, because there are underlying APIs for cells and data types.

I know that keyboard shortcuts can sometimes be more efficient than GUI actions. So when I sometimes tease those of you who don’t want to lift your fingers from the keyboard, please realize that I still have a feeling of horror about all those keyboard commands I had to type in the old days, and a deep appreciation of the fact that a GUI doesn’t disrupt my train of thought about what I’m doing, in the way that “extraneous” keyboard commands do. I appreciate the intuition that I can “physically” move and affect items on my computer screen, and “see” the relationships involved in an organizational structure.

But I’ll confess that I really don’t like using a mouse, as that does require removing my hand from the keyboard. I prefer a PowerBook or (now) MacBook Pro trackpad, which to me is simply an extension of the actions I perform on the keyboard.

48

I am kind of indifferent to the whole keyboard/mouse debate maybe because I just don’t perceive the differences for me. I also don’t have any great need to archive random bits of information because my work involves gathering a HUGE amount of information on a singe subject. To whit, I have a 6 gig collection of archive web pages, text documents, etc. I am unwilling to commit that collection permanently to any proprietary database because I have learned from hard experience that nothing stays the same forever and you always lose something valuable in exporting data in order to switch systems. Also, I need everything to be accessible to a desktop search (I use Foxtrot because Spotlight is horrid). So, to make is short, I store everything in its native format, html for example, with the reference metadata marked in various ways according to the different systems I have used. I then import data into my analysis program of the day which at the moment is DT of course.

As for Tinderbox, I think some people miss its real value. I see TB as a kind of “visual spreadsheet” which lets you organize data in a diagram format while the nodes contain the actual data itself. As an Investigative Researcher, I sometimes work on a project which is just to complex to organize in an outline form which leaves out DT. Such a project often lends itself to the TB visual organizing format. Also, TB’s scanning functions allows one to find connections in data that has fixed attributes which can be identified by the program. I would’t dream of using TB for simply storing data. That just doesn’t seem to be what is designed to do.

Anyway, I hope these ramblings are interesting to somebody.

49

I’ve just reread this thread and still have a question that may be too naive, in which case, apologies to all.

I’m a professor and, therefore, writer and information gatherer. Most of my gathering consists of notes, and many of these notes are less of the “general info on the web” type and more structured notes on books and articles. The main points of my notetaking are to 1) identify key passages and capture them; 2) comment on these passages in ways that are easily linked to the passages themselves; and 3) use these first two modes to generate original ideas. I have used Tinderbox for a number of years, and, while I share many of the general frustrations aired in this forum, it is useful for keeping precisely these kinds of nested links to specific sets of ideas. I have also used DT for about a year and a half, and find it useful as a general storage and retrieval mode, but less useful as a writing and notetaking tool.

So, this is a long windup to the question: is there something I am missing in DT? Or are there some fundamental style differences between tracking ideas and tracking bits of data and information? TBX is better for the former, for me, and DT much better for the latter. Are there other forum members who track ideas and arguments unrolling over multiple texts who have other opinions? And if you do take notes on ideas primarily in DT how do you organize these different layers of original text; notes on text; new ideas?

Thanks.

50

Hello,
I do similar kind of work have a complementary question: Is there a way to structure and organize highlighted text, documents, and links? Using DTPro over the last months has been amazingly helpful for organizing writing and the features that let you draw on the AI capabilities are helpful too, but too often lead to really spurious sources that take a long time to vet.

When I “collect” information it usually is for one or up to several purposes and beyond the folders, duplicates, and replicants, I can’t see a way yet to create a multidimensional index (e.g., index plus table of contents) to organize the information I have to help focus my searches using DTPro and DA.

thanks,
Francis

51

I don’t know if this is going to help because it is difficult to step into another way of working and thinking about that work, but…

It seems to me that what you are asking comes down to a question of the “metaphor” which represents the software in question. That is, having used both TB and DT, I would describe DT as an outliner and TB as a “visual spreadsheet.” I don’t have any fundamental problem with DT because, in the end, I write reports and those reports have to be structured and I can’t imagine doing that without an outline. As I said in my earlier post, if the material is too complex to outline in the beginning, I use TB or a drawing program to help find the structure but eventually it all goes into an outline. Once the report is done, I generally don’t use those “ideas” again but I do draw upon my data set which is archived outside of DT for the reasons I also posted about earlier.

So unless I am suffering from a failure of imagination, the bottom line is that if your work can be conceptualized as an outline of some kind, then DT should do the job. If not, then I guess you would would have to look elsewhere. It would be great if DT could incorporate some kind of non-linear metaphor where a store of information could be designated as a “node” of some kind ala TB but I have given up on looking for the Shangri-La or ogranizers a long time ago and am happy when I just find something useful like DT!

52

I also use outlines to begin writing, but have seen TB and DT almost exactly opposite from what you describe. I understand TB as non-linear (or really tree-structured) outliner. DT as intelligent database. Can you describe a bit more how you see and use DT as an outliner? Thanks.

53

I’d also like to know how DT is used as an outliner. This seems counter-intuitive to me - I struggle to see how I can effectively use DT as an outliner. Like lfriedla, I see, and definitely use, DT as an “intelligent database”. All my research goes into DT - absolutely all of it, notes, articles, PDFs, websites, everything. It’s AI makes it so easy to group related information together (I love replication) and to find information later that I don’t know how I managed before it. I can’t imagine writing my doctorate without it. Groups make it easy to manually find information I want, “See Also” helps find the unthought of links. And “Find” helps me quickly locate specific articles, quotes or authors. And I know that I am not yet using DT to it’s potential.

However, I don’t use DT for writing and don’t intend to begin to. I use Mind/Concept mapping software (Inspiration because of it’s ease-of-use and excellent graphics) for mapping out the major themes and core content, then export the resulting outline to a word-processor to actually write (I still use MS Word, partly because of familiarity but also because my university requires that I submit a MS Word version of my papers). I didn’t realise how visual I was until I used Inspiration, and am still learning how powerful adding images to concepts can be.

The power and utility of DT, for me, lie in its information management. Nothing else I have tried comes even close.

So how do people use DT as an outliner?

54

Papyrus, unfortunately, is unusable if you need non-Western unicode because it does not support other keyboards so far.