Consider window management, for example. If we're talking about using the keyboard for nearly everything, window management is one of the most fundamental parts of interacting with a computer. Linux has dozens of keyboard-based tiling window managers. You have only a few options on macOS, and they are all weak by comparison. Yabai is passable if you partially disable SIP so it can manage spaces, but it still has pretty limited functionality compared to herbstluftwm and bspwm (which is what it's mimicking) and feels hacky.

Similarly, cli tools are very popular on linux, and it is very easy to bind keys to them. Luckily a lot of tools work on both linux and mac, but this is not always the case. Replicating my global hotkeys has been an annoyance on macOS in some cases because of this. In linux, you have control of and multiple options for basically everything (e.g. bootloader, what you use for networking, what you use for notifications, etc.). In the case of notifications, on linux I can just bind a key to "dunstctl close-all" to close all notifications or "dunstctl history-pop" to show old notifications. My notification daemon is designed with scriptability in mind. If I want to have a similar keybinding on macOS I have to go through an arduous process of trying to find some up-to-date applescript snippet that actually works and isn't slow (couldn't find one that met either requirement) or learn to use some other automation tool. Unlike the WM case, I'm sure it's possible, it's just that macOS is just very clearly not designed with the use-case of keyboard for everything in mind. It can still support it very well, but with linux I can have it so that non-keyboard-friendly software just doesn't exist on my system. I don't need to workaround it.

It would be cool, but I just don't see any of this as necessary. What I use supports single-key shortcuts, modality, and modifier shortcuts, and I wouldn't need anything else. Emacs and probably vim support arbitrary key combos and some more advanced functionality with timeouts, but I don't use those features. Even at the global level, I don't use tap-hold on alpha keys or arbitrary combos. It seems like these extra features are primarily useful as workarounds for having a bad keyboard. Normal modifiers are good enough for me. There are already plenty of actions available using just single keys and modifier keys, and modifier keys are comfortable on a good keyboard (even on my personal laptop I have 8 comfortable thumb keys). "hold T for Toolbar and tap 1,2,3" seems like a comparably painful/inefficient chord, and I'm not using software designed primarily with mouse interaction and things like toolbars in mind.

> But it's not a binary, even apps that offer customizable key shortcuts, mostly offer underpowered versions nowhere close to what a keyboard remapper can offer.

I'm definitely biased in that I'm a software engineer and am not working with graphical applications like photoshop. 95% of my time is spent using tools that are programmable and intended to be used primarily with a keyboard or at least have first-class support for it.

> there is also a more limited browser-only, but cross-platform extension GhostText

firenvim can also actually embed vim in browser input fields regardless of OS. Vim is really the only option still if you don't want an external window though.

> Though maybe you can set a keybind that would instantly open a specially configured editor

Right, there are things like emacs-everywhere that can be globally bound and will automatically handle the "syncing." Probably there are similar plugins for other editors but nothing universal I'm aware of.

> All of them do, e.g., app launcher requires the contexts of the app state. If the app is in the foreground, it will hide it, if it's in the background, it'll show it, if the app isn't not launched, it'll launch it.

I have global keybindings for all my most used applications that support toggle/dropdown functionality. I don't really use enough applications to need a launcher that does this, but you can at least implement this using external context only. It would be easy to script something like this on linux, but it would be X11-specific (e.g. tdrop + maybe rofi) or window manager-specific.

> That would still fall short of Keyboard Maestro with its GUI where you can much easier find he available actions and stich them together

Only if you're a fan of GUI configuration. I'd much rather have no GUI at all and just good documentation.

> It's also the only way to make something great. There is no "native" OS/app development platform that matches that power, so it's very rare for an app to implement anything advanced on its own

> > You have some context available but only what is externally available > That's a very common limitation unfortunately

It certainly would be nice if all software was integrated at the level where you could access and work with internal state. It's why I'm using Emacs for nearly everything, but then you're stuck with whatever limitations/flaws the system has, so there are tradeoffs. If you like having options, it's nice that linux has so many Desktop Environment choices and dozens of really good window managers. By comparison, if you like tiling window managers, macOS is not great. Yabai is passable but requires partially disabling SIP or else it can't even manage/switch spaces. On the other hand, a lot of hammerspoon functionality is missing on linux. Linux will probably never have any good generic automation tools. X11 at least has some more advanced, generic tools like xdotool, but this is currently impossible with Wayland, and definitely nothing supports both. Anything more advanced than input like window management is wm/compositor-specific currently on Wayland, so if anything, linux is going backwards wrt generic keyboard/automation tooling.

The maintainer has been very responsive in my experience and would probably add more advanced conditions if someone wanted them.

> There are way too many interactions for me to list

Sounds like mostly hackarounds for non-keyboard-friendly software. I generally prefer to use keyboard-friendly software (of which there is no shortage on linux) and just make application-specific keybindings directly in each application. Anything else is inevitably a hack. Window management, browsing, anything related to text or programming, etc. all have good keyboard-friendly options. I have to use some stuff like excel and teams for work, but I'll just use the mouse more in those cases. I don't spend a significant amount of time dealing with applications like that though, and e.g. for Teams I'd try set up a matrix bridge or just use warpd before trying anything else.

Teams no longer supports linux, but for other apps that have non-customizable key shortcuts, it's definitely possible hack around that and create your own in linux. E.g. When I used to use libreoffice regularly, I set up modal keybindings that sent the corresponding chords to libreoffice.

As long as I can use my editor to type anything non-trivial, I'm good. For example, you have generic software like sketchyvim on mac to essentially use vim in any input field. I don't think there's a linux equivalent that lets you essentially embed an editor like vim inside any input field, but you can use a generic keybinding to open one and then have the text sent back when you're done.

A lot of the examples you list don't seem to need any app-specific or context-aware functionality (e.g. your launcher example). I don't have enough of an idea of what you actually want to know if what's available on linux would support your use cases, but in addition to the numerous basic hotkey daemons there are programs like keyszer that support application-specific keybindings ootb or hawck which is scriptable in lua and supports conditional keybindings (allowing you to script per-application keybindings). You have some context available but only what is externally available, i.e. unless the program provides some way to interact with it like a cli utility or ipc socket, all you're going to be able to do is send input to it.

Keyboard Maestro would be replaced by one or more completely different tools. There are programs that allow application-specific mapping. I've never seen the point of it, so I can't comment if any of them are any good. What are you looking to do? As someone who uses the keyboard for everything, my experience so far has been that macOS is much weaker at this.

In your case, I'm not sure why you would need to roll over to other keys from escape. My shift key doubles as escape, and I don't use any delay or checks at all since I don't rapidly type escape then other keys. In any case, you could try configuring your capslock key to only act as control if it being held fully overlaps with the next keypress. Not sure about keyd, but kmonad and kanata both support this (I'd highly recommend kanata). That may require changing how you type depending on the modifier, which is why more fine-grained overlap threshold checks can be useful.

Relevant github issues:

https://github.com/kmonad/kmonad/issues/351

https://github.com/jtroo/kanata/issues/128

Mu4e now uses Gnus' code for viewing emails by default. Mu4e's headers mode will look a lot nicer than Gnus' article mode though.

Gnus is massive and has a huge learning curve. It would be great if you are working with very active mailing lists. If you're dealing with a small amount of email, there won't be much benefit.

This is not really objective because you also have to take into consideration how intuitive/simple the UX is. Given 100 related actions, clicking through a logically grouped nested menu or multiple pages to do something is going to require more steps but be something anyone can do. Memorizing and using keyboard shortcuts for all those actions will be more efficient, but isn't something someone could immediately do. There is of course an in-between where you can have menus to display available keyboard shortcuts, but some (most?) people will still prefer a GUI/mouse-based approach. There is no one-size-fits-all approach.

The "separate package" issue is really a non-issue, especially in Clojure where you have a different namespace for every file. Some people prefer a separate package for every file in CL, but even with one-package-per-project, you only have to worry about your own code. Anyone redefining functions outside your package isn't going to cause any problems with your macros.