The current crop of operating systems features horrible support for (multi-)tasking actually.

Motto 1 for the new era - Using a computer should always be easier than not using a computer
Ted Nelson (1974)


How do we support a continuity of intention across a long task? In the context of the browser, how do we contextualize pages and interactions inside some abstract task? How do we clarify context? If the human is forced to multitask, can the multitasking tools be grounding instead of stretching us out?

The state-of-the-art solution to remaining on track is putting a post-it note on our screens to not lose track of the task we are carrying out.


“Thumb in a book” kind of digressions (interactions) are prohibitively difficult, with two results:

  1. The digression is postponed, and the user carries it out later, with less context (or not at all).
  2. The digression happens, but entails a fairly heavy context-switch cost, leading to mental fatigue.

    This also shows up in, e.g.:

    Endnotes in scholarly books are an abomination in normal times, but even more so when your only access to most books is online.

Why are endnotes harder to use on the computer?

⇒ State not explicitly accounted for is “heavy” and irrecoverable (https://www.commitstrip.com/en/2016/09/22/ctrl-control/ 🌐)

Tool-belt model

I miss the possibility to put a window/program/something in my toolbelt, so that it is accessible at all times no matter the context - say, a pop-up widget or a drop-down window (like Yakuake 🌐!). A drop-down terminal can technically substitute for all other tools, but when I’m writing, I could really use a drop-down Thesaurus, Wiktionary, Wikipedia…

Isn’t this kind of a larger-scale “forgiveness” principle?

The forgiveness principle states that user actions should generally be reversible and that users should be warned if they try to do something that will cause irreversible data loss.

II 🌐 - Forgives mistakes

Information which entered the machine through deliberate operator action shall never be destroyed or otherwise rendered inaccessible except as a result of deliberate operator action to that end.


Aren’t desktop “windows” actually a downright terrible way to interact with a computer?

The primary mode of organization - floating, rectangular, overlapping chunks of interfaces - seem to me actually hyper-narrow and specific in their use (i.e. virtual post-it notes).

(This also evidenced by most prominent softwares that utilized floating windows eventually backed away, like GIMP)

In most (if not all) other situations, one wants to maximize the possible screen estate (especially given the relatively low screen space available, see thoughts.interfaces.pixel-space) - hence the primary use-mode of tiling window managers.

(Someone noticed also: https://desktopneo.com/ 🌐)

Why not have the full arsenal of metaphors (split windows, tabs, trees of tabs) at the level of the OS and let it take care of it - allowing one to construct their workflow and interface set-up as necessary, minimizing context switches?


(Probably should be in a different note)

Screenshots, despite their ubiquitousness and power, still compress a wealth of information into a bitmap.

Projects like https://screenotate.com/ 🌐 are amazing, but still ultimately feel like the wrong approach?

Additional resources: https://www.ianbicking.org/blog/2019/01/overengaged-knowledge-worker.html 🌐