I do not have any RAM to share, sorry.

Economics simulation + Python needing 200+GB of RAM sounds preventable.

In your friend’s shoes, I might start asking for pointers over on the programming.dev Lemmy.

As others have said, a rewrite in a faster language like C or goLang could help - but my guess is there’s also ways to cut that memory need way down, while still using Python.

I was just thinking about this: more evidence of the Minecraft to Linux self hosting pipeline.

we’re maaaaaybe 3% of the market on a good day, so they say “fuck it”

So true. And worse than that, we’re probably also the 3% most likely to skip buying a game that requires anti-cheat, anyway. Many of us are famously un-friendly toward closed source code running with invasive permissions.

There is probably no way to opensource it without also making it easier to bypass.

I want to highlight this in case OP missed it. Your point here is critical.

Now I’m going to nerd out a bit about it:

To expand on your points above (for OP), there’s an impasse here between the anti-cheat developer and the distro developers.

The anti-cheat developer needs support from the distro developer to get their anti-cheat packages signed, to allow them to run in the kernel. Any package not signed by the distro developer that tries to run at kernel level will be treated by the OS as a virus. (Windows has this protection as well.)

Getting the code signed is pretty easy. The only requirement is sharing the source code, so the distro developers can make sure there’s nothing nasty in it.

But the anti-cheat developers feel that they need to never share their source code, to prevent cheating. In some cases, they have even have contracts that prevent them from legally sharing parts of their source code (if licensed from a third party).

That’s also not a problem. All they have to do is sign a binding contract for secrecy with every contributor to the distro, and then privately share their source code, and get it signed.

On Windows, that means signing a contract with Microsoft. On Mac, with Apple.

But on Linux, is just means tracking down and making separate agreements with a few thousand independent individuals…

So the technical solution is pretty simple: share code, get code signed, run in kernel.

But the contrasting needs of everyone involved make it unlikely on Linux.

Interestingly, an Anti-cheat developer who felt very confident that their code was unbeatable, could just publish it publicly, and get it signed and running quite quickly.

But uh… Most anti-cheat is also pretty low quality code, according to most estimations.

I’m enjoying the dedication to great defaults, by the Gnome team.

The issues I experienced this evening on windows were there by design.

That’s exactly what has kept me loyal to Linux. When I do have an issue, at least no one designed the issue on purpose to abuse me.

I’m also not a fan of being limited to certain cloud services (I use neither Google Drive nor Dropbox).

Good news!

Boox will run any cloud sync that has a recent Android app available. It’s just an Android tablet, at heart, and it’s already unlocked.

In particular, the local NAS sync client for Synology runs like a dream.

Edit: Oops. I missed that we switched from Boox to reMarkable, there.

I guess I’ll leave this since Boox sells eInk tablets that are feature matched to reMarkable, anyway.

But why?

Probably because having two separate dependency management solutions can lead to a lot of needless headaches.

And it makes particular sense for Gnome to switch over, since Gnome is focused on user space apps. Flatpaks should generally be more relevant and lower risk, long term, since they don’t require root privileges to install.

without some sort of clear ‘Linux Certified’

I think we will get there. Ironically, we might have ‘Steam Certified’ first, for alternate SteamDeck hardware.

I’m the meantime there’s premium brands like System76 and Framework who are making Linux support part of their brand image.

Needing ndiswrapper was bad, but Ndiswrapper was salvation (okay it was fixed wifi, which is practically the same thing, to me…)

Yeah. I’m confident in this change thanks to my equal discomfort configuring either SELinux or whatever OpenSUSE has. Hooray!

But joking aside, this sounds like a good thing and I appreciate all the folks who do understand that stuff and have worked hard on it.