Red Hat's source code for RHEL (Red Hat Enterprise Linux) was previously publicly accessible, even if you were not a customer. Now only customers may get access to the source code (which is allowed by the GPL since source code only has to be delivered to those who have received binaries generated from it). But there are Linux distributions who use Red Hat's publicly available sources to create RHEL "clones" (in quotation marks because they obviously don't pretend to be RHEL), except without providing the corporate support one would receive for being a RHEL customer. They do have community forums though.
The superficial issue is that those "clone"-distros would have to either purchase a RHEL license or apply to one of Red Hat's other programs to access the sources for their own distro. The actual issue is that Red Hat's terms for being a customer are that they'll kick you out if you use that code to redistribute your own versions of it (or, god forbid, even create a full distro from it).
Since CentOS proper was killed off years ago, many people who wanted a Red Hat compatible server distro but didn't want or need commercial support shifted their systems to the aforementioned other "clone"-distros, which are now in danger of disappearing because of that change.
Is Red Hat legally able to do it? Yes. Is it a dick move? Absolutely. Will it help spread the popularity of RHEL or other Red Hat distros? Absolutely not.
Uh. The relationship between CentOS Stream and RHEL is a bit murkier to me. I'd be lying to you if I said I fully understood how that code flow works.
For openSUSE the flow is "openSUSE Tumbleweed" -> "SUSE Linux Enterprise" -> "openSUSE Leap"
Everytime SUSE creates a new version/service pack of SLE (SLE 15 SP4, to use an example) the sources for that version are provided to openSUSE, and a new version of Leap is released (openSUSE Leap 15.4)
I don't actually work on Leap much, nor am I a SUSE Employee, so there are probably some minutae in that process that I'm missing, but that's the basic workflow.
It can't be retroactively amended, that's not how contract law works. The GPL isn't something you sign up for and you have to accept updates. It's a model license that people are free to use instead of writing their own. While a lot of people do license their software under "GPL version X or later", that is not mandatory and it would severely hurt GPL adoption if it were made mandatory in a future version.
At best that's extremely debateable. The GPL explicitly precludes placing any other restrictions on receiving the code outside of the ones in the GPL. A paywall to receive code from them is allowed. Terminating access for someone exercising the rights explicitly granted to them by the GPL sure as hell sounds like an additional restriction to me. The entire contract you have to sign to receive the code the first time most definitely is.
You did not mention the acquisition of Redhat by IBM in 2018. At the time, many observers were worried about what IBM would do with Redhat, and always wondered when IBM would go corporate crazy over a company with a model based on open source. So this decision crystalize the doubts that many people had at the time, that IBM would milk Redhat more, and this induce a kind of "Yep, of course they did go corporate crazy!" reaction.
Many people feel burned and don't want to invest more efforts in a company with this kind of methods (they did not even allow a grace period) and want to jump ship.
Have distro hopped over the years - most recently Manjaro to Fedora to Endeavour, but haven't found the one that's quite perfect for me.
That said, I'd make a few recommendations based on the person I'd be "marketing" to:
New to Linux, looking for polish: Mint
Mint is built off the well-known Ubuntu, polished a step further. It's in my experience the simplest to use and most generally polished of the Linux offerings. The community generally isn't as catered to power users, but if you care more about your time than about customization, I'd recommend Mint.
Looking for Stable/Modern, willing to jump thru a few hoops: Fedora
Fedora has come a long way over the years. It's far more stable, polished, and accessible than ever before. I'd hazard to call it my top recommendation, BUT, third-party software management and installation can be something of a nightmare. COPR is approximately equivalent to the AUR of Manjaro/Endeavour/Arch below, but at this time very obtuse and difficult to learn or work with. Some day you'll want a package that exists in COPR, and that day won't be fun for you.
Need apps you can't find anywhere else: Endeavour/Manjaro
Forget bleeding-edge packages and rolling release - the Arch User Repository (AUR) is hands-down the greatest feature on offer from Arch-based distros. The AUR is a repository of packages created by users that aren't supported by the main repos. If ever there's a time you need a piece of software and you can't find it anywhere else, the AUR's your best bet.
That said, I found/find both Manjaro and now Endeavour to be a little rough around the edges, and the consequence of rolling-release and bleeding-edge software is a system that isn't always working just right.
Looking to learn, straight into the frying pan: Arch
Same benefits and drawbacks of Endeavour/Manjaro above, but if you want to set up your system service-by-service, as lean as you want, Arch is there for you. A great experience if you just need an excuse to "try" putting an OS together piece by piece, even if you don't ultimately keep it in the long run.
Desktop Environments
The great DE debate. Nobody can tell you what's right and wrong here, but I have a few general breakdowns of the "big three". GNOME: If simplicity and elegance is your style. You sacrifice customization potential for cohesion and polish. KDE: Modern. Powerful. Usually polished out the gate. Can be a bit much if you're trying to tweak it tho. My personal choice. XFCE: Less modern, more friendly to lower-end systems.
Great write-up. Arch was my first experience with Linux on the desktop, and I learned so much. Would highly recommend it, even if you don’t want to stick with it.
I agree with most of the suggestions, but I would recommend staying away from Manjaro. They don’t seem to know what they are doing.
Their idea of releasing updates is delaying them from upstream (Arch) with little testing. Updates frequently break the distro. They let their SSL certificates expire way too many times.
RedHat creates a product called RHEL (Red Hat Enterprise Linux) that is a paid support product, mostly targeted at businesses (and things like Academia/Laboratories/etc).
At one point, there was a Wholly seperate product, created outside the RedHat umbrella, called CentOS, that quite literally took the sources of RHEL, removed the RHEL branding, and rebuilt it, allowing folks to "mostly" be able to use RHEL, without paying RedHat for a support contract.
In 2014, the CentOS Project/Product was "purchased" by RedHat, and then in 2020, RedHat decided that CentOS would no longer just be a "rebuilt" RHEL, but instead would become the development space for RHEL, called CentOS Stream. This made many people very unhappy, and they decided to start the Rocky Linux and AlmaLinux projects to provide roughly the same product that prior versions of CentOS had provided.
Additionally (I don't actually know exactly when), at some point, Oracle started doing basically the same thing that CentOS had been doing, and rebuilding the RHEL sources, and selling it, as "Oracle Linux"
So net effect of what this means, is that RHEL sources will no longer be publicly available at git.centos.org, and will only be available to RedHat customers (i.e. you must have signed up for an account/license with RedHat for RHEL). This may make things more difficult for Rocky, Alma, and Oracle, to provide the same "Bug for Bug" compatible product to RHEL.
Most of what people are upset about, is because they're willfully misreading the GPL (GNU Public License) which covers an awful lot of the RHEL sources.
The GPL requires that if you distribute software, licensed under the GPL, that you also must provide the folks that you distribute that software to, with the sources you used. It doesn't specify how you have to provide them, you could make them available for download, you could mail folks a DVD with all the sources on it, (honestly, I think you might be able to just print them all out and send them on dead trees, and still be compliant).
What most of the folks are upset about, is there is a clause within the GPL, that says something about providing the sources "without restriction on redistribution" or some such. And they view that RedHat can choose to terminate your license to RHEL, if you redistribute RHEL sources/software as violating the GPL. But the GPL cannot dictate business relationships. Redhat cannot stop one of their customers from distributing sources that they are licensed to have. But they are well within their legal rights to terminate that license, and provide no further access, if you distribute them. (i.e. you have an RHEL license, and version 1.0 of a library is covered under that license, you redistribute that source, and RedHat must allow that, but they're under no obligation to continue that business relationship, and provide you continuing access to version 1.1)
That's a rough rundown on the history. What does this mean for the average linux user? Nothing, really. Unless you happen to use Rocky Linux, AlmaLinux, or Oracle Linux. It doesn't affect Debian, or Ubuntu, or openSUSE, or Arch, or anybody else. RedHat will continue to contribute back upstream to projects like the linux kernel, or GNOME, or what have you, they will continue to sponsor and hire developers, they just will no longer be providing free and open access to the RHEL Sources.
It's not a question of legality really, but more one of an ethical nature. It sort of depends on you, as to whether or not you're bothered by RedHat doing this or not.
Thanks for the thorough explanation - What does this mean for Fedora OS? Isn't that maintained by Redhat as well? I have a fairly large fedora server base in my homelab and hope I don't need to redeploy it all!
Unless you happen to use Rocky Linux, AlmaLinux, or Oracle Linux. It doesn't affect Debian, or Ubuntu, or openSUSE, or Arch, or anybody else.
So, stupid question, but would Fedora be affected at all? I know that's related to Red Hat, but I'm guessing it's not affected since it's not based on RHEL.
It's not a question of legality really, but more one of an ethical nature. It sort of depends on you, as to whether or not you're bothered by RedHat doing this or not.
I'd say I'm bothered by it, but there's not really anything I can do about it. I'm disappointed the GPL doesn't have stricter rules regarding the distribution of source code though. I feel like it kinda defeats the purpose if sources aren't freely available to anyone who wants to use them.
The only thing I think you may have gotten mixed up here is that CentOS or other clone distros didn't remove the branding. Red Hat did that themselves in thier repositories that were used in the clones.
If I'm remembering correctly, in the very early days of Centos and the like, that was the deal that Red Hat had struck...you don't use our trademarks/branding and you can have access to all of our source. Most likely so that Red Hat wouldn't get endless support tickets without pay if something went wrong on a clone package.
There is a legal question. It would have to be litigated to be resolved, though. The argument is that when RHEL threatens to cut off future business, they are placing a restriction on redistribution. RHEL would argue that it's a restriction on future business, not on current redistribution, but who can say what a court would make of that distinction.
It's true that RHEL does not have to continue a business relationship in general, but the point here is that they need to follow the GPL when making relevant business decisions.
Of course harassment is never okay, but I'd say when it comes to GNOME, this is not surprising. GNOME developers have been so hostile towards both users and other developers for a long time. I'm not saying every single person associated with the project does this, but it is pretty common (e.g. here and here ). Of course the GNOME devs don't have to accomodate everyone, but it is a common theme with the project to remove features despite user backlash and also to close bugs as WONTFIX often without good explanations as to why, even when there are pull requests for fixing the problem.
I am simply avoiding the project, since there are enough good alternatives.
Definitely agree with this sentiment. Another aspect for myself -- When I tell Linux to do something, it fucking does something. No questions, no obfuscating, no disallowing in order to "save me from myself". I have a significantly increased sense of control over MY hardware with Linux.
It is frankly unfathomable to consider going back to the telemetry / spyware laden corporate controlled systems of the mainstream OSes.
@Zeppo Yeah I might try the Dual Boot system to begin with to save me time working out VMs but I think the long goal would be running a windows VM on Linux tbh.
If you must dual boot, have totally separate hard drives, and the choosing between Linux and Windows should be done in the UEFI boot menu, not GRUB menu. Windows can render Linux unbootable otherwise, requiring a rather complicated rescue. Windows would ideally not have any chance to see the Linux hard drive while booted.
An external SATA SSD in a USB enclosure is cheap these days.
Yeah, Windows likes to overwrite the MBR with no warning as if that’s perfectly fine. I’ve always wondered what combination of carelessness, incompetence, interface streamlining and competitive malice is responsible for that. It’s also ridiculous how in 2023 there’s still no native way to read Linux filesystems from Windows.
@Zeppo That does sound ridiculous especially as I wouldn't mind running linux through Windows, I know it wouldn't work as well but I know I would do a lot of my less complex tasks through there such as browsing the internet as I could game on Windows or Linux as I don't really mind if I'm VMing one
@Zeppo The goal would be to use Linux as my base OS in the future. As I look through the many comments from everyone I am now re-evaluating some of the things I feel I 'Need' in my set up such as my Sync Cloud Storage, Instead I hope to move over to External Hard Drives.
Knowing Adobe is a huge part of my setup does hurt the idea of moving to Linux at the moment but I will have to find if there is a workaround other than VMs in the future. People mentioned GPU Passthrough on the VM which would help a lot, It's just not only learning how to run a VM but how to do the pass-through on it.
But yes I do wish to move to Linux in the future but maybe baby testing the Distos I think I might like on a VM might be the best way to step into this rather than going into the deep end straight away.
Sure, installing Linux in a VM would be the fastest and easiest way to test it out and start becoming familiar! You could try a few distros/desktop environments and see what you like without having to deal with reformatting or real hardware.
I do the external usb ssd for my windows drive and recommend it endlessly. I use the windows drive for music production and the maybe 2 games in my several hundred game library that don't work in linux. 100 bucks for a 1tb samsung external ssd, and wintousb to make it functional.
@technologicalcaveman Thank you never have heard of a WinToUSB but will really need to look into it as it sounds like a great way to be able to run Linux.
It's a good tool, pretty easy to set up too. I personally recommend not connecting to internet when setting up windows 10 so that you're not forced into either signing in or signing up for a windows account. I did that and about a year later on the same install, it's all good still.
I'm sad to hear about this, vim is a great creation and it's sad to see someone go. I didn't know anything about Bram but vim is a significant tool I use daily.
@nlm Thank you, these will come in handy and I'll have to test it out on my test rig (It's not the best but if I can test out a program on it, Might as well.)
For starters, I just want to clarify that this isn’t a shot at OP, more so this is a criticism of the linked article.
So the article can be summarized as such:
Ransomware is a problem
There are Linux versions now.
Some comments about Linux being on 3% of the desktops now. Not sure what the point of this was… since the author then says that Windows is still 80%?
Linux servers are a much bigger/viable target, because the Internet runs on Linux, including financial sector, databases, “always on” systems, etc etc…
Ends the article with generic best practices like “deploy modern EDR” and “proper backups”.
All in all, this article contributes very little in terms of informing the reader or making any sort of original or insightful claim.
Apart from Ubuntu/Fedora (which are Snap/Flatpak heavy), I think you would be OK with any Linux distribution. I have a Intel Atom N270 and 2GiB of RAM happily running Debian Bookworm and KDE (with an SSD) your talking about something with far more power.
For me the considerations are as follows.
RAM
You've listed 4GiB of RAM, looking at my PC now (Debian Bookworm, KDE Desktop, 2 Flatpaks, Steam Store and Firefox ESR running), I am using 4.5GiB of RAM.
2.9GiB of that is Firefox,
~800MiB is Steam of which 550MiB is the Steam Store Web Browser.
~850MiB is the KDE desktop
Moving to XFCE or LXDE would help you reduce the Desktop RAM usage to 400MiB-600MiB, but you'll still keeping hitting memory limits unless you install an addon to limit the number of tabs. Upgrading 8GiB in would resolve this weakness.
I get by on the Netbook limiting it to 3 tabs or steam.
Disk Storage
You've listed 500GiB of HDD Storage, this means you want to avoid any distribution which pushes Snaps/Flatpaks/Immutable OS because the amount of storage they require and loading that from a HDD would be insanely slow.
Similarly I would go for LXDE or KDE desktops, both are based on creating common shared system libraries so your desktop loads one instance of the library into memory and applications use it. As a result such desktops will quickly reach 1GiB of RAM but not increase much further.
Also moving from a HDD to SDD would give noticeable performance gains, the biggest performance bottleneck as far back as Core 2 Duo/Bulldozer CPU's was Disk I/O.
GPU
The biggest issue will be the 710M, I don't think NVidia's Wayland driver covers this era so you'll be stuck on X11. Considering the age of the GPU and the need for the proprietary driver, personally I would aim for Debian or OpenSuse the long release cycles mean you can get it working and it will stay that way.
From a desktop perspective, I would install KDE and if it was slow/tearing I'd switch to Mate desktop.
KDE has some GPU effects but is largely CPU drawn, it tends to look nice and work
Gnome 3 choses to use the GPU even when its less efficient so if it doesn't work well on KDE it won't on Gnome.
Mate is Gnome 2 and works smoothly on pretty much anything.
Cinnamon is Gnome 3
XFCE is like Mate is just works everywhere, personally I find Mate a more complete desktop.
Open source software usually doesn’t do any kind of tracking or telemetry. Sometimes it is there but then it’s usually opt-in (off by default unless you change it). Most Linux distros contain a huge amount of open source software. So all the code is in the open (which means usually no shady things going on, because a lot of eyes are looking at that code) and there’s often not even an incentive for the developers to gain money directly from the users, except through donations. So if you really like some open source project, please do contribute or donate to it.
So in the end it doesn’t really matter which distro, they all will be MASSIVELY more privacy friendly than any closed source OS like Windows, OSX iOS, or proprietary Android, although to maximize your “chances” so to speak you can go for a purely community-run distro not backed by any company (Ubuntu is backed by Canonical, Fedora is backed by Red Hat, OpenSuSE is backed by SuSE, these are the three big distros with a commercial background. I’m not saying they do violate your privacy currently, but they at least have a greater tendency to do so, because data gathered can be sold, so it might be a business incentive for them. Ubuntu sent users’ search queries to Amazon in the past to gain some more money, but the community outrage caused them to remove this anti-feature afterwards again. Fedora is currently proposing to introduce opt-out (on-by-default) telemetry, it’s not decided yet, but it’s a bit worrying still. So you see, such distros might not be the best choice for “maximum” privacy. But compared to Windows or OSX they’re still magnitutes of miles ahead.
For community-run distros, you can check out e.g. Debian or Arch Linux, or any distro based upon them. For Debian specifially, I recommend running its “testing” branch, because it’s more up to date. Don’t worry about the label - it’s still rock solid stable, because Debian has very rigid testing requirements. They test more and longer than probably any other distro, which means Debian “stable” is very well tested, but also quite outdated. To alleviate that a bit, you can use the “testing” branch. You could even use the “unstable” branch for even more up to date packages, but there’s at least a chance that you get some package dependency problems every once in a while or so, so not recommended for a beginner. Debian is also quite easy to get into nowadays, though maybe not as easy as some of the Ubuntu-based distros. Linux Mint (normally Ubuntu-based) also has a Debian-based edition these days, and Linux Mint is a great distro for beginners. Arch is hard to get into but great for modern desktop usage or gaming because it’s always super up to date. You can also check out EndeavourOS, an Arch-based clone with easier installation. Or just use any distro, really, it’s not that big of a deal currently. They all behave quite well. Mint, Kubuntu or Fedora are good starting points for beginners.
If you have to use a public computer temporarily, there’s really only one choice for a private usage in that scenario: Tails. Put it on a USB stick and use it whenever you’re on an “unsafe” computer. Tails ensures that ALL traffic will be routed through Tor so no one on the local network (or the web, for that matter) can sniff out your data transfers, among other things it does to enable anonymous computer usage (e.g. it leaves behind no logs, doesn’t save any info about your previous sessions, and so on).
It’s a bad idea to recommend Debian Testing for people who have never run Linux before. Obviously people can do whatever they want, but the whole point of Debian stable is that it is stable, and the whole point of Testing is that it’s not.
Based on the ~2 videos I’ve seen, the newly released Debian 12 stable might actually be good for newbies without being noticeably out-of-date. Thanks to Flatpak etc, new software versions can be installed / updated easily without compromising stability.
Linux
Hot