If you’re concerned about this coming from Meta, you haven’t been paying attention. This is not to excuse Metas responsibility in the enshittification of the internet, but as a hyperscalar they of course have been pushing computing technology for a long time, and many of it openly available. CRAM will probably be upstreamed into Linux and I assume is therefore already licensed under the GPL. Meta Research seems to be operating in a much more open and useful way than the rest of the company.
Cool, pervert ram, just what everyone asked for
Meta can CRAM this functionality into mark Zuckerbergs ass
ZStandard is also developed by a Facebook. And its truly amazing. What’s wrong with their CRAM development, if everyone benefits from it for free and is Open Source?
As a user of Z-RAM, I am excited
Nope, never going to use this now. I don’t trust Meta to walk a dog, certainly not with a core os component.
But why not? What is the fear here? It’s not like Meta has any control if you install or enable the feature.
If it is good and open, engineering is engineering. Good engineering came out of Meta before, that all of us use daily, like the zstd compression algorithm.
Save the moral outrage for all the other shady shit Meta imposes on humanity, god knows there is a intercontinental cargo ship load of it to choose from.
If you genuinely realise (and care) that meta is shady, there is no merit in supporting the company.
Moral outrage is something that people with decent morality have. Apathy is something that zombies have.
Meta has funded all sorts of open source infrastructure, for a long time. If you want to avoid their funded contributions… TBH, I wouldn’t even know where to begin.
I got bad news for you if you use the modern web
Name a few things, I’m curious.
Ever heard of a tiny little thing called React? GraphQL and relay? Presto? Folly? Zstd? Hack? RocksDB? Btrfs?
PyTorch is crazy big in ML/AI and has been for a while now.
And specifically in Linux: https://engineering.fb.com/2015/10/05/open-source/improving-the-linux-kernel-with-upstream-contributions/
And that’s not even mentioning things that aren’t just their tech, like open compute and funding foss and the Linux foundations, crap ton of open source projects, testing, bug fixes, and miscellaneous contributions… Let alone all the tech that tries to compete with their tech; all the alternatives to react et al that are ‘next gen’
Like it or not, the facebooks and googles of the world have contributed significantly to things you probably interact with every day.
Heard of btrfs, don’t use it. Nor the rest.
0 out of 10.
As to the Linux kernel at this point everyone has contributed something. That would be like being mad at air for having been processed by Zuckerbergs human form cooling unit.
Zstd / ZStandard is pretty good.
Now you’re just not arguing in good faith here. You asked for contributions, there they are. I use Btrfs and it’s great. (There’s lots of actually valid reasons to hate meta, let’s focus on those)
The criteria was “that all of us use daily” not “that you’ve heard of”. Scroll up, it’s right there.
You use btrfs, I don’t, therefore the “all of us use” qualification fails, let alone “daily”.
A few notable ones?
https://engineering.fb.com/2018/10/30/open-source/linux/
Btrfs
https://itsfoss.com/news/react-foundation-launch/
React
PyTorch (if you’re into the AI thing)
I suspect, without evidence, that you’d probably find some Meta engineers contributing to the Linux kernel to some degree, so you may have to rip that out of your system.
0 for 3, is that it?
Impressive, building your own kernel without Btrfs. I admit, you’re better than me.
I don’t get this comment. You asked for examples, and they gave you.
Embrace, Extend, Eliminate?
You mean like ZStandard from Facebook too?
That is not how this works. EEE can be said about window’s wsl, but here they are contributing directly to the kernel. That is why linux is that good, these big corporations profit billions by running linux, and getting it sligthly more memory efficient with current ram prices makes them profit even more. Fortunately the license forces them to contribute the code back and we profit from the same optimizations they implemented.
Fortunately the license forces them to contribute the code back
It doesn’t. It would only do so if they distributed their modified kernel to others, rather than simply running it on their own servers.
They chose to contribute their changes upstream anyway, presumably so they aren’t burdened with having to adapt them to every new kernel version in the future. Fortunately, we stand to benefit as a side effect of that decision.
I dislike Meta as much as anyone, but sometimes a cucumber is just a cucumber.
Here is your cucumber:
People who dislike meta don’t promote meta.
I dislike Meta. But I like their work on ZStandard and use it myself. Best compression algorithm since 7zip. It’s extremely fast and can be used in may applications for various use cases. If that is the standard we can expect, then I’m happy what they do on CRAM. We only benefit from this.
Let’s hope GPL v2 helps
