And then I accidentally found this article on ACPI debugging, which references the memo written by Bill Gates in 1999:

One thing I find myself wondering about is whether we shouldn’t try and make the “ACPI” extensions somehow Windows-specific. If seems unfortunate if we do this work and get our partners to do the work and the result is that Linux works great without having to do the work. … Maybe we couid define the APIs so that they work well with NT and not the others even if they are open. Or maybe we could patent something relaled to this.

What. The. Heck.

This is insane… Isn’t it like the textbook definition of lobbying? I wasn’t expecting to find a whole conspiracy while trying to fix my Deck, perhaps the memo is a hoax or something, but this all just lines up so naturally. If it really was his plan, then he succeeded.

Given that the the memo was submitted in court as evidence in a 2002 case, Comes v. Microsoft, it’s probably real. If anything, it didn’t succeed enough. It probably would have been possible to lock Linux out entirely, but by 1999, there were already too many Linux and *BSD x86 server deployments. Couldn’t ignore them. Had to make it just kinda shitty rather than battening it all up.

Server Linux has exactly the same ACPI problems and manages fine there.

Or stretch out the twists in the individual wires. That will also cause signal issues.

IIRC, cat5 cables are rated for 50lbs of force on them. They’ll technically hold a lot more than that, but you can’t guarantee the twists will stay in spec.

Red power receptacles. Was this at a hospital?

Have you seen videos or pictures that have dark sections, and there’s “banding” where there’s a noticeable difference between something black and something very black? Like a sharp border where it’s obvious the conversion process from the camera to your screen didn’t fully capture a gradient of darkness?

That’s due to the process not being able to handle darker areas compared to very bright areas. It’s not enough to have an HDR display; the whole chain before then has to support it, as well. When it’s done, not only does it get rid of banding, but finer elements in darker areas can pop out and join the rest of the scene.

Copyright, yes. And a lot of this is corporate history rather than the legal portion.

SCO Unix was mostly dead before then (not fully dead, just smelled like it). They were never the most popular Unix vendor to begin with. Caldera–a commercial Linux distro–had bought them out, and that’s when the legal trouble started.

All those old vendors tended to have one specific thing they were really good at. IIRC, the thing for SCO was that they could load up hundreds of users on a single box on 1990s hardware. No small feat when the traditional Unix model needs to fork() a process for login/shell/whatever.

In a sense, NextStep is the only one of the old Unix vendors to still have a significant install base.

DejaVu Sans Mono. Open source, good Unicode support, clear distinctions between characters (“iI1oO0”).

That’s where I was a few years ago, and then I switched back to proper Linux. I was only keeping Windows at all for games, but then most of the games I played started working fine on Linux (thank you, Valve).

Plus, I tried doing some TensorFlow stuff with CUDA (Nvidia) GPU acceleration. In theory, you can do it in pure Windows, but nobody has bothered trying to do that. You’re on your own if you try it. The usual way is to do GPU passthrough to WSL. There have been three different ways to do that over the years, only one of which currently works. If you happen to Google a page that tells you one of the wrong ways, there’s a good chance you’ll need to reinstall to get it working the right way.

Using pure Linux for this stuff is no problem. Just use Nvidia’s server drivers instead of gaming drivers. All the AI datacenters are using Nvidia GPUs on Linux, so Nvidia is highly motivated to make this work. Someday, Windows might be as easy to use as Linux.

The one you toggle into the switches on a PiDP-11.

Glad we ask his opinion first.

Does RMS even have any code in the Linux kernel?

It’s for programmers who need their Imposter Syndrome amplified.

This is alright if you only know the basics of vim and then learn further from within that environment. If you’re already an intermediate to advanced user, then the keybindings between VSCode and Vim tend to interfere with each other. You’ll have to relearn how to do it.

Having more than two browser engines out there would be nice for standardization reasons.

OK, I can see the whites of your eyes.

SVG, unironically yes. There’s a few times where I found a library or WYSIWYG editor making some strange choices for its SVG output, and I had to fix it manually.

I think there’s a good reason for that. If you’re not as concerned about resource consumption (Emacs used to be called “Eight Megabytes and Constantly Swapping”, back when 8MB was a lot), then there’s no reason to avoid even more complex and resource intensive IDEs. People who wanted a complex editor, but in a relatively small footprint, stuck with some variant of vi.

Thus, vi found a stable evolutionary niche. It’s a tardigrade.

I’m having the opposite problem right now. Tightend a VM down so hard that now I can’t get into it.

Did Gentoo ever untangle the NP-complete problems in its package manager?