Our great sponsors
-
InfluxDB
Power Real-Time Data Analytics at Scale. Get real-time insights from all types of time series data with InfluxDB. Ingest, query, and analyze billions of data points in real-time with unbounded cardinality.
-
WorkOS
The modern identity platform for B2B SaaS. The APIs are flexible and easy-to-use, supporting authentication, user identity, and complex enterprise features like SSO and SCIM provisioning.
Swap is indeed supposed to prevent this AFAIK. You can though try some tools like EarlyOOM and see if it helps : https://github.com/rfjakob/earlyoom
If you want completely monolithic & uniform experience built for your specific needs that no one else has on Linux but you, I don't think you'll be happy, ever.
> consistent titlebars and ui elements across all applications
You can get fairly close to this if you use gnome or kde specific distros.
But you're generally asking: please don't be open source. Please don't have a bazaar of ideas. Please build me one big cathedral. You ask here is antithetical to the purpose of whom you are asking. On most general purpose distributions, users probably ought to end up having multiple different UI elements.
> complete keyboard remapping such that macos keybindings work everywhere
I actually like this idea a lot, because it suggests a certain system-wide malleability layer that, at the moment, doesn't exist. I'd be interested to see how people thought we might tackle this.
Folks could make a custom Linux distro that pre-configures each app to be Mac like. I think that's the best chance. But jeeze it seems like an unholy crusade to support a very specific niche, a niche not known for participating & giving back & relishing what we are & do.
You could use something like https://gitlab.com/interception/linux/tools to read certain Mac key-combos or what not and rewrite them. But what would you rewrite to? I don't have a good sheet of what it is you'd be asking for or wanting to just work.
My spitball idea for how we'd really fix this: I'd like each app to register with a dbus service all of the "actions"/commands it can do, and allow rebinding & activation of the actions over dbus. Maybe even actions actually are just dbus methods, but annotated somehow, to describe their hotkeys or to give them human friendly names? Anyhow, whatever the impl, there could be a central "hotkey manager" that could see all keyboard bindings & let us top-down manage them. There'd need to be some way for "Save" in Kate to be combined/grouped together with "Save" in Firefox, somehow, for this to be helpful. Managing this namespace of actions would be a terror of a problem, utterly absurd, in my view, which implies strongly the difficulty of the ask here I'm trying to respond to, but I actually think it'd be a pretty noble & cool effort. In part because of what it relates to:
System command apps. Tools like dmenu/alfred/albert/quicksilver are meant as general top down interfaces, are often scriptable/extensible/deeply configurable to allow fast access & control of a variety of actions. By recognizing keybindings as what they are: actions/commands, and suggesting that the "actions"/commands of an app get bubbled up to the system layer & get managed there, we just make these top-commanders more powerful. There's also an extreme parallel here to voice-agent systems, like Chrome Assistant, Alexa, Siri, where apps present actions & the system is in charge of taking user input and translating it into actuation; they too are directory systems of actions, rather than having each app in isolation.
> ability to copy and paste into or out of a terminal without fuss
Copy/paste just works for me? Not sure what the problem statement is here, and/or what terminals you've suffered under.
> Desktop linux is not being held back by system usage. If anything, we need to stop caring about that for a while and focus on quality of life / ergonomics.
100%. My main personal laptop is 4GB. I run sway which is low resource consumption, but in general resource consumption seems like a huge non-factor to me. In general, Linux isn't going to win by being more conservative. Pining about resource consumption is self-rewarding, self-gratifying: one feels zealous & virtuous, like you have the true cause amid a fallen world & are the path of the defender. But IMO it's mostly detracting & abusing the good & necessary & vital suffusion of creativity & possibility into the world. The scope of consumption is not that bad. And we have the important task of figuring out where to go still ahead of us: I'd rather be conservative once we have better ideas of what works, at any resource budget, & hone back down from there. Rather than forever dance around this maxima/minima we're on & tune for what we have.
I'm also unimpressed with this article in general. Showing the amount of memory mapped in seems incredibly uninteresting & indicative of nothing. Amount of data read has some correlation with start time but loosely: if Gnome is reading 1GB sequential (it's not but for example) while KDE is 512b reads randomly (it's not) but half the size, you'd probably still want to pick Gnome.
I am spoiled on macos, where cmd is the primary modifier key. So cmd-c and cmd-v are for copy/paste, and ctrl-c still continues to work in a terminal. I understand the need for ctrl-shift-c in a linux term, where ctrl-c needs to remain for signaling processes. I just wish I could globally remap this. Some of the hacks exist for this (https://github.com/rbreaves/kinto) but it's a leaky abstraction and not fool proof.
https://github.com/serprex/nobox would make for a good baseline, it uses about 12K resident memory
It was my desktop environment for years until I switched to wayland & sway
In regards to point 1 about keybindings, I've used evdev to get something similar to Mac(my alt is CMD).
https://github.com/csumtin/cua