-
Notifications
You must be signed in to change notification settings - Fork 0
Conception
We have a few nouns:
- A human.
- A project.
- A task.
- A context.
- A lens.
A human makes and changes tasks, in lenses, subject to contexts, towards projects.
A human may be on many projects.
Tasks are part of one project.
You pass through different lenses on the site, either making, taking, detailing, or completing tasks.
Insert Ben's drawings here.
Brainstorm is perhaps more accurately dubbed "Braindump": just get the things out of the user's head. Yes, all of them. They shouldn't put any effort into making the thoughts clearer or more definite or anything else, that can come later. Right now we are just concerned with getting everything out of their head with as little friction as possible. The goal is a tight "have a niggling ought-to feeling" => enter into system => feel more comfortable and stable that things are under control.
The big thing about refine is that it corrects the necessary errors of the brainstorm page. There need to be 0 barriers or hesitations to using brainstorm, or people will second-guess and ease-of-use plunges. So after brainstorm, users (especially untrained ones) are likely to have malformed lumps of tasks instead of the bricks planning needs.
I imagine a deck that feeds into a CSS well with 4 exit nodes from the well: Delete, Defer, Delegate, and Do
In the well: extra fields (like maybe a short name, tabbables to tags/due dates, etc?) pop up to let the user hone in on what they meant. I suspect the background values of the fields will be the most important in teaching on-the-fly. Finalizing the task somehow moves to "where does this go?" mode.
Do: starts a timer and will prompt with a "done?" button. Just do it now. Defer: Puts the card in the "refined" version of the deck for planning. Delete: Puts the card in the garbage box Delegate: Starts the other-inbox/email flow for a task, leaving it in the owner's deck but marked for another person.
I think spatial metaphors are the big thing here, because it will be hard to get people like my mom to understand anything else.
So overall, a deck of cards is presented, and the first one is queued into the well. The user can edit the additional fields that appear to provid details of what the the potentially-malformed task actually means. That done, the user can make the task to-be-done by them, or three variants of making it not a problem (giving it to someone else, deleting it, or just doing it.) The next card queues up, and the cycle repeats. When the deck is empty, there is a prompt to move to planning.
Plan was originally conceived to handle dependencies of tasks, but observed us cases (with Allie) have shown users putting subcomponents of a major theme in sequential layers, ex:
- Upload
- Format
- font
- color
- JS
- Download for retest
Becomes a DAG of
- layer 1: 1 node
- layer 2: 1 node
- layer 3: 3 nodes
- layer 4: join back to 1 node
** Befores & afters in a DAG might be sufficiently powerful that we should intentionally underdefine the MVP to see how people use it**
Transform a DAG into an executable pile, befores presented before their afters, like a deck of cards in your pocket. Because of all the prep work that went into making the cards executable, there should be no rethinking of the work at this stage; just look at the top card and do that.
Overview is hardest because we don't know how users will want to use it yet. I could imagine it being useful to track "What the hell was I actually doing this week," but that is just me. Conjecture about use cases:
A barchart with days on the X axis and total execution time in a stacked bar on Y or count of tasks executed
A gantt chart of the last week, with each task as a rectangle, each day as a row off the Y axis, and time during the day on the X
A barchart of tasks on X with time open on Y, or days til due (smaller in front as it is scarier)
A calendar with stacks of cards on each day? Toggle between done and generated?
Each task in a separate row, with start time and end time, sorted by duration?
Histograms of task time by tag?
We should experiment with our own tasks to do research.