Reccomend a next task

is a wonderful article by Aaron Swartz that details the failings of productivity software.

The key feature he finds missing is the ability to reccomend a next task.

The infromation needed to intellgently reccomend a next task is broken down and includes time estimates, priority, and context.

While this is about programming it was really enlightening to me, reccomending a next task is exactly what I want from my paper based planning system too. It might be small scale (you have put bread in your grocery cart, now get tea) or larger scale (a major project is due next week, to be on schedule you have to finish your reserch and citation checking today).

I am still absorbing the implications of this insight, but wanted to share.

Syndicate content

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

So a priority grid/checklist?

I think I see what you're after. I dabbled a bit in project-management, and we would always dump project tasks into a 2x2 grid. High-priority tasks were in the top row, low-priority in the bottom, low time-estimates in the leftmost column, and high time-estimates in the right. Tasks are listed in each box, or if you're doing planning on something like a whiteboard, tasks are jotted down on stick notes and placed in the grid. The higher the task (closer to the top) the sooner it should be done. You could set up one grid for each major context (home tasks, work tasks, errand tasks.)

I don't know that such a system would be good for the minutia of remembering to buy tea :-) but I have found it useful to get my head around programming work for myself -- what features are the most important, and what do I have time to implement right now?

the difficulty of suggesting a next action

Interesting article - not at all what I was expecting. He seemed to focus on the fact that productivity software should be able to pull in actionable items from other pieces of software: emails, bug tracking databases, project management software. The thing is, unless there is a standard API for listing actionable items from all that stuff, you're stuck with the very solution he complains about people suggesting in the Postscript: writing plugins to import this information.

Now, as far as software actually recommending a next action, I can see that going two ways. It seems like it would be really easy, especially in a GTD application, to add functionality to say I'm @phone and have 15 minutes, what tasks can I do? Order by importance and due date...

On the other hand, let's say you put visit with a good friend from college you haven't seen in years on your list of stuff todo. Time estimate: all afternoon. Context @somewhere away from the house or work. Due date: none. Using just about any computerized algorithm for picking what to do next, this would NEVER happen. Unless you put it high priority or due soon so it will show up.

That right there sounds like you're decided what to do next and defeats the whole purpose of this functionality! I remember from reading Getting Things Done, people are good at intuitively deciding what to do when they have a list of what they can do. They are just bad at remembering what to do, and when to do it (buy flashlight batteries at the store, not when the power goes out and the batteries are dead).

So - in my opinion: software (or paper solution, even) that lists what you can do given your time and location: good. Showing that list ordered by due date and priority: good. Having the software decide for you what you should be doing next: bad.


I don't really want

I don't really want softweare to do this, but I do find the question is really useful in thinking about my planning. I tend to overcomplicate things and asking myself if something will help me select a next task is really helping me pare down and focus my planner.