"Ajar source" is definitely the funniest take I've seen on this. Kudos, Bryan.
> In other news, Richard Stallman's head has exploded.
I'm honestly not sure about this one. If I'm recalling Stallman's takes correctly, I think he's fine with "Ajar source" as long as the source is open to Red Hat's customers. I remember him saying something along those lines when explaining how he expects developers to make a living-- something along the lines of it being okay to charge money for your product, but then your customer should have [some list of freedoms].
RMS has always been clear that selling Free Software and only distributing the source to people who bought the software from you is fine. But I don't think he's ever suggested that somebody selling Free Software in this way might threaten to cut their customers off from future updates if they ever exercise their right to redistribute the software they bought. Redistribution is a key part of "[some list of freedoms]", and this "you can redistribute this version but if you do you'll be cut off from future versions" scheme is not something I've ever heard him endorse.
*> Redistribution is a key part of "[some list of freedoms]", and this "you can redistribute this version but if you do you'll be cut off from future versions" scheme is not something I've ever heard [RMS] endorse.
He likely wouldn't endorse it, but I don't think he would condemn it as non-Free. After all, if a company wants to "fire" a customer, they're allowed to do so, for whatever stupid reason they like.
As far as the customer's freedoms are concerned, as long as they're aware that by redistributing they're pulling the ripcord on the relationship and starting (or switching to) an unfriendly fork, well, that's just an oddly formal version of the process.
But Ubuntu are if anything worse in their own ways.
So if there is any RH to Ubuntu traffic, I don't expect it to stay.
Anyone considering Ubuntu should probably just use Debian. Not that Debian is the ultimate answer to all needs, just that "if you were considering Ubuntu".
But really I don't know what to suggest these days. Alma & Rocky have a strategy for the moment but who knows if RH will figure out a way to break that at any minute.
That leaves what? Arch? Nix? Gentoo? no. Maybe just everyone makes their own based on Alpine or something.
I think we can probably count on Debian staying true to Debian, so that's a good choice. Besides that, if somebody was using Red Hat and wants to continue using a distro with corporate support, SUSE Linux Enterprise seems like the obvious choice to me. Also OpenSUSE Tumbleweed is great for home/hobby use.
I actually had chosen opensuse for my own employer for almost 20 years (holy cow, but yeah from about 99 to about 2018) and then switched to Centos due to my feeling that the engineering at Suse had gone down the toilet.
I don't know if Suse got better in the last few years but obviously Centos 7 was an unlucky time to decide to commit to the next 20 years of Centos future.
But opensuse was great for most of that time, and at it's worst, better than Centos now. Suse at least haven't tried to pull either RedHat or Ubuntu shenanigans. They're looking pretty good now!
Why are people still referring to "Red Hat" instead of IBM? Back when that sale was announced the talk was that Red Hat's culture wouldn't change, they'd be operating independently from IBM, so it made sense to still speak of Red Hat. Well there has evidently been a dramatic shift in Red Hat's culture; do people still believe this "independent subsidiary" stuff?
That's what it's come to? I have been using Ubuntu for about 15 years, but with snaps still starting up as slowly as they do in finally switching once support for 22.04 runs out.
You don't have a choice any more unless you go very out of your way to remove the scaffolding on a fresh install.
Which of course I have done long ago on my laptop, so problem solved!
Then the other day I'm trying to spin up an aarch64 vm in qemu to test some things. What I'm testing is a gui app, which launches a url in the default browser when you press one of the buttons (github.com/bkw777/mainline , it has a button to open the upstream ubuntu mainline-ppa web site to see the extra info about available kernels) I just want to test that everything works when arch is not amd64.
So I grab the latest aarch64 ubuntu image and give it 2 cores and 4 gigs out of my laptops i7-1280P w/64g and gen 4 nvme
I need the vm to be able to run my dinky little app, and firefox, but otherwise doesn't need a full desktop running, just the X apps installed and I access it by ssh -X.
This thing was ridiculously slow, and the culprit was all snaps and snapd. Yes emulating anything in qemu is slower than not, and emulating a foreign arch is worse, and only giving it 2 cores and 4 gigs is not much either. And yet, other distros besides ubuntu ran in the same environment just fine, no worse than a cheap vps. The ubuntu installation was also fine, after removing snaps.
But the act of removing snaps was AGONIZING. It takes a stupid amount of time, like several minutes, just to uninstall one item for some steps. Uninstalling firefox in particular is just insane. What in the world can it possibly be doing for so long just to uninstall? It wasn't just stuck. The whole time it was running it was outputting occasional messages about unregistering things. This is in an otherwise fast and idle machine. A laptop but a good one, plugged in to wall power, in full performance mode, with good cooling. the qemu disk image was configured for raw not cow, the underlying ssd is a fast wd sn850 2T on a fast gen 4 pcie m.2 slot. Even if I had some sort of mis-configuration that crippled it, this does not change the fact that other systems ran fine in the same conditions, and even ubuntu ran fine after ripping out snaps and snapd.
And of course it was ridiculous to spend that much time crafting the install of a throw-away test vm. Current Ubuntu basically totally and utterly failed to be good and useful for this very ordinary random task. I only did it because the app I was testing is actually Ubuntu-specific. (well, debian-based-specific, but most users are Ubuntu and that's what needs to be tested)
I am tempted to reproduce the whole process and capture it all on video just to show in a way that can't be ignored or victim-blamed just how utter UTTER shit Ubuntu with snaps is.
It cannot be overstated. There is no level that qualifies as hyperbole. The fact that people have stupid high powered pcs not doing anything else with all that power such that the ridiculouslness is more or less hidden in day to day casual desktop use is no excuse at all.
My car will still drive more or less as ok as before in most day to day situations even if I fill the back seat with cinder blocks. That does not mean that if a car manufacturer decided to ship their new cars with 10 or 20, or even just a few cinder blocks built in that it wouldn't be utter shit and no one should buy those cars.
I believe the author is providing an amusing viewpoint in this post rather than a serious one. After all, the final sentence is: In other news, Richard Stallman's head has exploded.
> In other news, Richard Stallman's head has exploded.
I'm honestly not sure about this one. If I'm recalling Stallman's takes correctly, I think he's fine with "Ajar source" as long as the source is open to Red Hat's customers. I remember him saying something along those lines when explaining how he expects developers to make a living-- something along the lines of it being okay to charge money for your product, but then your customer should have [some list of freedoms].
Maybe the telemetry does it though. ¯\_(ツ)_/¯