Conversation

· Mar 4, 2021

In cogsci, Marr suggests 3 levels at which a system (eg vision) can be analyzed: computational (the fundamental problem being solved), algorithmic (how it’s solved, abstractly), and implementation (hardware details). It’s an interesting taxonomy for analyzing tools for thought!

Show this thread

Inboxes only “work” if you trust how they’re drained. From a queue-processing perspective: the departure rate must be below some threshold. Inbox Zero “works” by aggressively increasing that rate (via defer, delegate, drop)—blunt, but ensures that departure rate > arrival rate.

This tactic requires you to make a decision about every item in the inbox. Maybe fine when queue is small, but explicitly deferring a task imposes an emotional cost, possibly unnecessarily: “inbox zero” is only necessary if the arrival rate *always* exceeds the departure rate.

If the arrival rate is variable and sometimes sits below the departure rate, you can still handle everything in a reasonable timeframe. A more ideal mechanism would ensure that wait times remain tolerable without insisting on a 1-day cycle time.

Spaced repetition can lower the stakes around destructive inbox-maintenance operations

The system should also consider the number of decisions to be made: I’d rather make fewer decisions but tolerate a longer average wait time. One way to implement that might be exponentially increasing deferral actions:

notes.andymatuschak.org

But what I’m saying here is that thinking about this design problem as a queueing theory problem is much more generative than just brainstorming up a cool UI mechanism, because it gives you a way to *reason* about the space of mechanisms, e.g. in terms of average wait time.

A lot of this is around an insistence on ephemerality. You click on a link in Twitter & you have to read the article RIGHT NOW because 1) it’s modal so you can’t do anything else until done, 2) no *history* so you can’t go back to it & 3) search sucks so you cant find it tomorrow

I feel like it’s silly to have to come up with “actions” for the user to take to organize their information. I shouldn’t have to ponder whether closing a tab or de-escalating it is the right move. Information systems should be fundamentally non-destructive and searchable.

Gmail ironically got this right — don’t organize your mail, search is good enough that you can always find it later! We live in this depressing duality where every third party tracks every link I visit, but *I don’t get to track the links I visit*.

If closing a tab had the same “consequences” as hitting back in Gmail, I’d never worry about it again. It’s purely a a visual action I’m taking, not a destructive action that then gets bandaided with “undo tab close” or “easier ways to bookmark” or other anxiety-inducing ideas.

Replying to

7:13 PM · Mar 5, 2021TweetDeck

This resonates, but people still seem to feel a great deal of anxiety around their Gmail inboxes, because they still implicitly represent a queue of tasks they intend to perform (with some implied SLA). It’s just that (since they don’t organize) now the tasks may be hard to see…

Replying to

True, maybe one of the distinctions is that self-generated “casual todos” (e.g. background tabs of stuff you want to read but don’t want to bookmark) may lend themselves to self organization vs. “external todos” like email where other people are effectively dropping tasks on you.

Olli Tiainen

@olliten