Braindump: Managing Documents II

Braindump: Managing Documents II

It’s a dreary Wednesday morning, but my spirit is lighter because my backlog of immediate to-do items is very short. As a result, I’m feeling like tackling one of the back-burnered projects on my Trello board! However, as I ticked through the list of candidate projects, I felt a gnawing discontent at the required effort to gather and organize the source materials just so I was ready to start the real creative work.

This has always been a sore point with me, and when I was in grad school I had stumbled upon a paper as part of my research on user interface design that echoed my feeling of discontent: most of my time was spent doing dumb stuff before I could do the actual work. I’d forgotten which paper it was, mistakenly thinking it had been something Douglas Englebart had written; it was actually J.C.R. “Lick” Licklider and his 1960 paper Man Machine Symbiosis that had struck me. It is a fabulous paper, and it was incredibly inspiring to me back when I was defining my philosophy of software design.

Anyway, in section 3.1 of Man Machine Symbiosis, Licklider writes about logging just what he did as “work”, and was dismayed at what he found. He summarizes his findings as follows (emphasis mine):

About 85 per cent of my “thinking” time was spent getting into a position to think, to make a decision, to learn something I needed to know. […] Throughout the period I examined, in short, my “thinking” time was devoted mainly to activities that were essentially clerical or mechanical: searching, calculating, plotting, transforming, determining the logical or dynamic consequences of a set of assumptions or hypotheses, preparing the way for a decision or an insight. Moreover, my choices of what to attempt and what not to attempt were determined to an embarrassingly great extent by considerations of clerical feasibility, not intellectual capability.

The reason I bring it up is because I’m facing this issue from a slightly different direction: document management. In April I wrote about the pain of managing documents that were numerous in both type and location, and I’d found it irritating that I had to spend so much energy re-establishing my “working context”. This is very much related to what Licklider was talking about with regards to “85 percent of ‘thinking’ time” going into clerical/mechanical work before he could do the actual thinking. There’s got to be a better way, right?

I know, I know…how long does it really take to open a few directories and launch some applications? It’s probably about 30 seconds, tops, going by the clock. For me, though, those 30 seconds are like running through a minefield, distractions waiting to detonate whenever I open a new window, every step of the way. I have a terrible memory, probably because I don’t like having to remember things in the first place, so my initial few seconds are spent overcoming that resistance. I try to recall exactly where the files are, and where I left off. Then, I have to somehow load those files without having ANOTHER thought while I’m waiting. For me, that is quite difficult; although my short-term memory is terrible, my associative memory is really good. EVERYTHING I see immediately reminds me of SOMETHING ELSE, often with considerable detail. So, every step I take to load all files, applications, and web pages is fraught with the potential for distraction. I’d like to reduce that as much as possible by reducing time and mental effort to “load the context” so I can work. I’d love to be able to quickly switch between projects at the speed of thought.

How might this work? It might be like git checkout <branch> for the brain, to use a software development analogy. A real-world analogy might be having dedicated rooms for each project, with every tool and resource exactly where you left it next to the giant project on the main table. To switch projects, you just walk to its room, and everything is there waiting for you.

That’s what I would like my document management solution to do.


  1. Eurobubba 8 years ago

    As I mentioned in comments on your earlier posts, I’m engaged in a similar quest. Lately I’ve been playing with a Mac app called Beanflows (see — no affiliation!) for GTD “project support files” — documents related to specific projects that I’m actively working on. Beanflows basically provides a special-purpose viewer and tools to manipulate a filesystem sub-hierarchy where I can keep project resources. Combined with a “fetch alias” AppleScript I wrote to pull in links to apps and other files I don’t want to move from their permanent homes, it provides a simple but handy staging ground that allows me to avoid the distractions of the full filesystem.

  2. @Fishpatrol 8 years ago

    If you wanted to take it to an extreme, maybe you could create Mac user accounts for each project. Leave files, browsers, whatever open and fast-user-switch your way between projects. And hope you never have to restart?

    From the brief excerpt you quoted, I’m not sure what you include in “getting into a position to think.” In the rest of the post you talk about attention/distraction, but it seems like getting into position includes a lot more (arguably) necessary work.

    For work, I write a weekly tip/marketing email. That often includes a brief brainstorming/editorial meeting, looking up information, writing an awkward first draft (which gets shared internally if it’s really awkward to identify a better direction), looking up links, revising, and finally managing any images and text styling.

    A lot of this feels like getting-into-position work, but the final email won’t get written without all that positioning (if that’s how you categorize it). Sure I’d be happy to fire from the hip and hit the target–but that never really happens. For me, accepting the messy process made it easier to get on with the work instead of getting distracted by the messiness or frustrated with the lots-of-steps-iness of the task of writing this email. $0.02

  3. Author
    Dave Seah 8 years ago

    Eurobubba: Beanflows sounds very cool! Thanks for sharing that!

  4. Author
    Dave Seah 8 years ago

    Fishpatrol: (I love your handle, btw) User-per-project is a genius idea. Have you tried it? Maybe I could do something like have a separate user for personal, work, and maybe goofing around. One issue might be shared resources (certain key Excel files I use), but I like the idea of saving context this way.

    Regarding getting into a position to think, the Licklider paper I references the gist of I was thinking, but let me be a bit more explicit.

    As Licklider notes, he found that he spent a lot of time—the bulk of his time, by far—acquiring data, managing people to graph the data, assembling graphs, and then finally looking at them to interpret them, which takes mere minutes to assess. The interpretation is the actual “work” that produces something new or useful, though it was only possible once all the meticulous gathering/graphing work was done. In his own review, he found 85% of his time went to the busywork. And here’s where he thought that computers could step-in and help alleviate that: hence, the idea of man-machine symbiosis, where the computer could handle the clerical stuff and man could focus on answering questions based on the data. That was the dream back in 1960.

    Fast-forward half a century, and now have computers that allow us to get our busywork done more rapidly, but there is also a lot more of it to go with the greater levels of detail required. Maybe Google Search is the closest we have to something that handles the tedium, along with the work that Wolfram has done with Mathematica and the Computation Knowledge Engine. But there’s still a missing level of integration, I think, between “knowledge capture/curation of a variety of disparate data types” and “creative synthesis.” My gut assessment is that it’s still very rare to see in software design, because the focus is on “one task at a time”; a typical Content Management System is a good example of a well-structured problem that isn’t very good for creative synthesis. Scrivener, the writing tool, is by far a better tool for synthesis because it allows you not only the ability to capture data, but to COMPARE it side-by-side. This is what Evernote fails to do, which is why I don’t use it. There are also tools like Tinderbox from Eastgate software, which I haven’t actually tried, but are pushing in this direction.

    As I type in this text box, I’m struck by how easy it is to flow in my typing. That’s because I’m drawing on past memories and experiences, so they are already “gathered”. I am also writing in text, which is by nature a linear medium that also tames my thinking as a byproduct. The text itself is my stream of thoughts, my typing captures my attention with very little friction since I can touch type, and the text is also the product that you will read. I can easily go back and edit what I wrote. I can’t, however, fiddle with a lot of styling controls, which would be distracting as they take my hands off the keyboard. Instead, I can use inline formatting with Markdown, which semantically is “emphasizing” rather than “applying bold format”. The former keeps me in the writerly thought stream; the latter is a production-minded aesthetic styling that is more indirect.

    My ideal for document management is to get as close to the experience of having everything I need within mind’s reach. These are projects that require a lot of learning, trial and error, and evaluation to make the simplest steps forward. I’m with you on the “accepting the messy process” in principle, but in spirit I find it aggravating that there are few tools that work with the way I think while also keeping me pointed in the right direction, augmenting my weaker mental functions with superior notation tools and other forms of scaffolding.

    I like your example about the weekly tip/marketing email you write, which makes me realize that scope might have something to do with it. The kind of projects that I’m thinking of are fairly abstract and ill-defined. In this specific case, the project is “make a better creative workflow”, which encompasses:

    • Capturing workflow issues (accomplished by writing these blog posts, which live on my website)
    • Creating a list of the issues, with as much detail as possible
    • Grouping, categorizing, and connecting the issues on a meta level
    • Synthesizing hypotheses about each issue and sets of “related” issues, identifying the key factors that are believed to matter the most
    • Creating a conceptual map of the hypotheses, to itemize and also relate each one to shared concepts.
    • Investigating, summarizing, and citing existing research from relevant fields of cognitive science, psychology, software design, organizational development, and who knows what.
    • define the set of criteria that tests the hypothesis
    • design a set of interactions that can be implemented with existing software or paper prototyping.
    • design software and develop software: selecting the environment and libraries, learning to use them, manage sprints to implement features the correspond with the conceptual hypotheses, test the software to see if it does anything that we’d hoped
    • Manage software release, deployment, installers in source controlled process.
    • Involve outside testers, gather tester feedback with issue tracking software. Handle issues, organize and categorize them, and modify existing concepts based on the observed patterns.
    • Communicate findings to the community
    • Repeat everything until a stopping point suggests itself, and some conclusions can be made, or something better comes along.

    Each of these steps has its own set of documents, which may be images or video, references, resources, raw data, access keys, email conversations, blog posts, paper books, etc of multiple types that ideally are shareable and synchable between the computers that I use. And this is just one project that’s on the backburner.

    Having written that all out, it occurs to me that maybe you’re right…I should just accept the messiness and just do the best I can. But I do wish I had a better tool that would help capture, compartmentalize, and help me recall pertinent information in support of what I’m doing.

  5. Eurobubba 8 years ago

    First off I should say that my own toolkit is very much a work in progress, but you might find it helpful to break down “document management” into more specific tasks and consider using different tools for each one. I’ve ended up using several applications that seem similar on the surface and have a lot of overlapping functionality to manage (mostly) PDFs for different purposes.

    I could probably get away with using just Evernote, DEVONthink, or Yep as an “anything bucket” AND a reference filing system AND a project support filing system AND a financial records repository AND a personal research library. But I find it helpful to use each one (along with Beanflows, Mariner Paperless, Scrivener, and various other apps that could all be placed under the broad “document management” heading) for a specific purpose. Besides letting me leverage each application’s particular strengths, I find this also helps me maintain focus and clarity when doing the work.

  6. @Fishpatrol 8 years ago

    I haven’t tried a multi-user project setup. Either I haven’t needed that level of compartmentalization or I just don’t want to. :) If you kept files in Dropbox, it might be able to sync across user accounts with enough speed that you wouldn’t notice that it’s working, but you couldn’t keep the same file open in two locations and make updates that would sync across. You’d probably need software that synced in the cloud instead of working at the file level.

    From what you’re describing of Licklider’s position, I don’t agree with him on many counts. Yes, drawing graphs would be tedious and time-consuming, but to me his examples are all real work because they’re sites for course correction and participation beyond the task. Gathering data is a slog, but when you read the answers, you see how the questions are being interpreted. You understand the data differently–better–because you’re closer to it.

    I’ve worked on a data project for a year and a half, tracking TV show prices in iTunes. I went about it in a fairly dumb way, both because I didn’t know how to use the advanced tools Apple provided and because I wanted to see the data and understand it better. I’m doing the whole back-end with Excel. After living in that data, I know so much more about it and what needs to be done to represent it well. I don’t have time to do all that on my own, and I’m shutting it down, which is really disappointing. But I understand it very differently than I would if I’d simply directed their data feed into a relational database, like they advise.

    A frequent question in that project was, how much effort is it worth to make a process more efficient? I’d often need to turn a database table into a fairly complex list in HTML. Enter concatenation! Mix text strings with cell references and it’s quickly done. But if you’ve copied down that formula across 200 rows and that day only has 30 rows of data to report, the formulas keep calculating. To keep them from showing, you’d have to add IF tests and work around Excel’s quirks so that the HTML column only has publish-ready code. It could probably be done, but I’ve stuck with copying the 30 rows in that column by hand. It’s a little slower than command-shift-downarrow, but the formulas are less complicated, and there’s very little time involved. It’s okay.

    There are other parts of the process that have been begging for improvement. Some of them would require moving to a different CMS and structuring data differently. That was the work hit the red-line for me. Just too many pieces to work on by myself. Put in a couple hundred hours of work to eliminate a 3-minute daily task, by itself, is a very expensive purchase. (Also I was trying to help those darn users, but let’s not get distracted by them.) The daily updating is a grind, no doubt about it. But the updates that got published were accurate. I can say that much.

    Back to the concatenating example, that’s a question I try to ask: Do I need to improve the overall process or just deal with a few edge cases? Sometimes it takes slogging through a process for a while before deciding it needs improvement. (The weekly email is what finally got me using Markdown on a daily basis.)

    But systems can be a trap, too. I do like a plain text document, but sometimes that’s too linear, too neat. Sometimes what I need is paper to draw on or index cards to notate and arrange. There are digital tools that metaphorically replicate these physical tools, but they often have to work Really Hard to get 80% of the way there. Between improving the process/tools and getting a job out the door, I at least sometimes choose to get the thing shipped. :)

    Maybe an interesting text would be to limit some stage of a project to be non-digital. If you need an image, print it out. Resources? Print them out. You write out what needs to be written with a pencil. Spread things out on an empty table, or tape them up on a wall or the tri-fold poster board used at science fairs. Maybe that would help spot parts of the process where the tools are working against you and a tangible mess is more efficient.

    Your “make a better creative workflow” project looks very interesting as the subject of a book. Are you writing a book? Maybe you want to write a book? I think for me, my focus on process tightens when I’m the only person working on a project. I have to keep everything running. If any change or improvement will get made, it’s because I do it. If there’s a new tool that would make the job easier and I don’t know how to use it, I’ll have to brute-force through or abandon it. No help is coming besides the help I build for myself. I did that with TV Bobber for a year and a half, until I couldn’t.