P.S., If I had my way at all, any IoT devices that end up in my home contrary to my wishes would be confined to a local network physically disconnected from the internet.

Alas, these days they probably come with a built-in WiFi chip configured to ride on whatever nearest clueless newbie-run open WiFi network it can find. 🤦 Y'know, just so they can phone home and provide "crucial" data to improve user experience. Oy.

Under Long Term Goals:

That's just... wow. Just imagine the day when, for the sake of "convenience" and "integration", this nasty piece of work starts connecting to your IoT devices and collecting data from them -- from your refridgerator to your internet-enabled toaster oven, your security cameras, (y'know, so that your computer can "help you remember what you had for breakfast" and offer to make the same thing tomorrow by sending a command to your toaster) and sending them to some anonymous server somewhere out in the anonymous internet.  And suddenly, some dude in a basement halfway across the world knows what you ate for breakfast and what color pyjamas you wore last night. As well as your list of contacts and who you last texted.

How does anyone in their right mind not scream bloody murder at this piece of nastiness??!

I for one am not looking forward to switching to Wayland. From what I've been reading online, it's a crapshoot. Well I could be wrong since I've never actually tried it yet, but I'm certainly not a fan of people shoving it down my throat when they haven't yet convinced me why it's superior (I don't think it is).

Set the environment variable LESS=X to force 'less' to stop clearing the screen upon exit.

Anyway, 790K LOC is totally insane for pid 1. The only place this kind of nonsense is even remotely reasonable is on Windows. 😜

I'm glad I took the leap and switched to Devuan, and restored some sanity to my system.

I am Ohm of Borg. Resistance is voltage over current.

A more accurate measure is the number of tokens in the code, because LOC can be biased by empty lines, comments, different formatting conventions, etc.. But even that isn't completely objective either, since it changes depending on language and what kind of abstractions are used. Though the general order of magnitude of LOC is pretty indicative of the complexity of the code.

But yeah, a million lines of code in systemd is like what on earth is it trying to do 🤨 And do you really want code of this complexity run as pid 1.

Lines of code isn't a very accurate measure of anything. What matters is the complexity of the algorithms used and how much work the code does to perform a given task. If a task is complex the code obviously needs more work and resources to arrive at the result. But given a particular task, we can compare how many resources are required by two programs in performing the task. The less resources the better, obviously.

There's also of how many features the program has, and how many of them are actually required and how many are redundant or superfluous. Kinda subjective, but if you have a specific usage pattern then a program that offers lots of features unrelated to your usage then obviously it's not minimal relative to what you use it for. If a feature isn't used for the majority of use cases, we can say that it's redundant and the program is bloated and not minimal.

Check if your environment has set the $MORE environment variable. That could be one source of implicit options. Also check if your bash profile has aliased 'more' to 'more' plus some additional options.

.bashrc, .bash_profile, would be places to look for this. Remove the options or set your own preferences there.

Ratpoison's code is about 20k lines, and compiles to a 180K binary.  So it looks like Stumpwm's code is more compact; the larger binary size is probably because it's written in Lisp and therefore comes with a full Lisp interpreter embedded.  It's also much more extensible than ratpoison, since lisp, being lisp, can literally rewrite itself at runtime.  But ultimately, the exact binary sizes aren't that important; what matters it the functionality to size ratio.

Wayland's premise is actually not bad.  There are a lot of things about the core X protocol that are clearly dated; there are many things that were important to display servers back in the day that today are completely irrelevant.  Color palettes being one of many examples.  2D drawing primitives that almost nobody uses because these days most programs render directly to buffer instead, or use the GPU's 3D rendering functions.  Font management API that is also mostly useless because these days people use the likes of libfreetype (which renders to bitmaps and completely bypasses the core X font APIs), not to mention that the font management is stuck in the bad ole days of 256 characters per charmap, with many hacks to workaround its limitations so that Unicode could work.  In short, many entrenched APIs could use an overhaul to better suit modern devices and usage patterns.

The execution of Wayland, however, leaves some things to be desired. Not caring for rare use cases that X supported is one of them.  Pushing the new API down the throats of people who haven't found reason to migrate is another -- this is the same problem that the likes of systemd, etc., suffer from.  An overly narrow focus on desktops and neglecting other legitimate use cases like remote clients, while being sold as the be-all and end-all of display protocols -- that just leaves a bad taste in the mouth when you try to use it thinking all the old apps would work as before, and instead hit the brick wall of reality.

They could have done this in a way that did not alienate existing users, but instead, they went and committed the biggest blunder one could when it comes to software engineering.

Stumpwm is written in Lisp. I'm not sure how lightweight that would be, but my first inclination is to guess that a C app is likely to be more compact. But that's just a guess, can't say for sure without actually measuring it.

Yeah, IIRC stumpwm is the successor of ratpoison, I think one of the devs may even be ratpoison's dev. Supposedly it's better than ratpoison, but I haven't tried it yet, I'm relatively happy with ratpoison as it is.  Besides, I'm developing my own WM in the same minimal vein, mainly for my own learning, and to try out some hacks I've been meaning to try.