For those who are out of the loop, like I was, here is how I'm understanding this...
"CentOS" tracked behind RHEL, and is considered more stable? Regardless, CentOS 7 is the last supported major version, and it's reaching EOL soon and will not be maintained. This sounds like a push to get those who use CentOS (free) to move to RHEL (paid) to stay in the long-cycle ecosystem.
"CentOS Stream" is a flavor that tracks ahead of RHEL, is less stable, and is really meant for those who develop on RHEL.
Since the original CentOS is no longer supported, you had new flavors come up (specifically Rocky Linux and Alma Linux) designed to fill the void to track behind RHEL.
However, it appears they rely on the open-source RHEL for their builds. It sounds like this is going away now, and the only open-source linux RedHat is gonna publish freely is CentOS Stream.
Please feel free to correct me if I'm wrong or missed something!
Backports are important for long term support. Say, you have some-pkg-1.2.5-83.el9.x86_64, CentOS Stream will have some-pkg-2.10.32-3.el9.x86_64 (with extra features, enhancments, security fixes), while customers want just CVEs to be back ported to some-pkg-1.2.x
This creates problems for long term support, as it requires more resources, who understand the code base for some-pkg in order to apply selective fixes. Right now, if one knows how to build rpms and how to integrate them is enough to become a clone of downstream RHEL.
I think, only Oracle can pull off this feat, as they can put more people to work to solve this problems. 1:1 compatibility becomes hard, as 1000+ packages need to be taken care of.
It has been cited by, among others, Bradley Kuhn of the Software Freedom Conservancy, which is probably some of the highest praise I've received in my career to date.
I suggest reading that and then if you have further points or questions I will be happy to try to address them.
> customers [...] are under the terms of a contract, which overrides the GPL license of the code itself
Do you have a source for this claim? Why would the contract be legally stronger than the license?
GPL prohibits further restrictions on rights of the code recipient. Contractual terms are sometimes found unlawful/invalid by courts. I guess this question can really be resolved only by a court.
Wow, that is pretty defensive/offensive and dissonant with your post above. You wrote in your article:
> [...] as far as we can see, the Hat is acting perfectly in accordance with the terms of the GPL [...] The key point being is that to obtain those binaries, customers – as well as developers on free accounts – must agree to a license agreement and are under the terms of a contract, which overrides the GPL license of the code itself.
And later here:
> I covered my take on the situation at length [...] if you have further points or questions I will be happy to try to address them.
So, if you were just reporting, and you're happy to address questions, my question is - who did you report there? Who thinks contract overrides the license? Is it you yourself, is it the Register team, or somebody else?
I think that if using that claim at all, you should have reported that this claim is pretty controversial, there are many discussions on the internet that show this.
> Wow, that is pretty defensive/offensive and dissonant with your post above.
Is it? Oh. Er, how?
> You wrote in your article:
> [...] as far as we can see, the Hat is acting perfectly in accordance with the terms of the GPL [...] The key point being is that to obtain those binaries, customers – as well as developers on free accounts – must agree to a license agreement and are under the terms of a contract, which overrides the GPL license of the code itself.
And later here:
> I covered my take on the situation at length [...] if you have further points or questions I will be happy to try to address them.
> So, if you were just reporting, and you're happy to address questions, my question is - who did you report there?
Wow. That is a sweeping request, and while I suppose it's a fair one, I can't directly answer it. I read everything I could find on this subject, from the statements of the various companies involved, the discussions on here, Lobsters, LWN, Phoronix, and multiple mailing lists including Fedora-devel.
I did not keep records on every page and every comment, so no, I can't answer. Not "am unwilling to"; I am unable to.
> Who thinks contract overrides the license?
Red Hat does, as far as I can see.
Red Hat has a lot of lawyers; IBM is even bigger and has more. RH is very careful indeed about exactly what the law says. It doesn't matter how many angry FOSS fans don't like its actions: it, as part of its professional due diligence, had to check first, before it acted, to make sure its actions were legal.
My understanding is, in summary, this:
The Red Hat and SUSE business model is quite simple, and yet quite widely misunderstood.
The companies do not sell software.
The companies sell support services. When a customer buys the support services, that gives the customer access to support, documentation, etc. and it also gives them access to the software, as well as to ongoing future updates to that software.
In other words, maintenance.
The software is GPL. You get it, you get the source. You can do what you want with the source, so no rights have been taken away. But if you distribute that unmodified source, then the company that provided it to you is perfectly at liberty to cancel your contract.
You don't lose anything. You don't lose the software; they can't take it back. You keep it, and you keep any documentation you have as well. But you don't get updates any more, and obviously, you also don't get support.
What you bought is a maintenance contract, and what you lose is the maintenance contract. You did not buy software, and you do not lose software.
Any customer contracts do not affect customers' GPL rights. They only affect the customers' rights to what they paid for: support and maintenance contracts.
I see a lot of entitled angry people who haven't put real thought into what's going on and are angry because something they felt they were entitled to has been taken away... but they were never really entitled to it at all. Realisation of that stings.
But that does not grant any more rights to them.
> I think that if using that claim at all, you should have reported that this claim is pretty controversial, there are many discussions on the internet that show this.
Absolutely. It is highly controversial. However, that doesn't mean that the large numbers of angry people are right. Numbers do not confer correctness.
Millions of people do not believe that human activities are causing the planet to heat up. It still is heating up very fast, and human civilisation will probably collapse very soon as a result.
It doesn't make any difference how many people disbelieve it, or how fervently their belief in whatever bogus nonsense they have picked up from the fossil-fuel industry. They remain just as wrong, and they will die just as soon as a result.
So, do you think your article is neutral reporting, or personal view/commentary? It seems to me it is the latter. That's OK, but then don't pretend you're just reporting and not playing a lawyer. Your last summary is pretty engaged and one-sided.
> I see a lot of entitled angry people who haven't put real thought into what's going on and are angry because something they felt they were entitled to has been taken away... but they were never really entitled to it at all.
That is pretty funny, because it works both ways, depending on your attitude to Red Hat :)
In addition to reading what people/organizations wrote since the announcement, it's eye-opening to read also past sources on the Red Hat attitude, especially how Red Hat tolerated CentOS, then embraced it and supported it [Jim Whitehurst quotes], then killed it, and now that they themselves have caused original centos rebirth (alternatives like Rocky and Alma), they've decided to make problems and even call the people doing useful work for the user community "freeloaders".
Jim Whitehurst quotes [1]:
"CentOS is a derivative of RHEL that works for people that want a stable release without support,"
"If they don't see the value of our model then at least I'd rather have them on something similar to RHEL."
"We don't view CentOS as a competitor. It's almost a complement,"
"Red Hat is great for production, but there are so many new innovations with OpenStack... There was nothing in the Red Hat family to support that. CentOS is moving faster and fills a gap in our portfolio of offerings."
The old management got the culture and did not attack the community. The present management is the opposite, and I think it's dumb.
The position that "they can't do this, it's GPL!" is risible. It's stupid: it means those saying it have not thought about what a corporation worth tens of $billions has to do before such a move.
Entitlement makes people stupid, though. Ask any person from a minority.
However...
> The old management got the culture and did not attack the community. The present management is the opposite, and I think it's dumb.
Analysis should involve also sources and argumentation. You stated a contentious belief as an obvious fact without source and argumentation. That's more like a commentary.
> The position that "they can't do this, it's GPL!" is risible. It's stupid: it means those saying it have not thought about what a corporation worth tens of $billions has to do before such a move.
That's your analysis?
No it's not risible.
Commentators do not usually claim this is their official legal position. It's an understandable emotional reaction to IBM/RedHat turning their back on decades of established mutual understanding with the community, which was that clones are fine.
What's risible (or sad) is that some people, on both sides, including you, think their position is obviously correct and the other one is obviously wrong. None of the legal questions here are obvious.
From a legal standpoint, the contract vs GPL issue is contentious, and if you do not see why, then you probably did not came across enough of various sources on the matter. There is a documented case in the past where Red Hat violated GPL as they threatened to revoke support to a customer using the GPL code if they won't pay royalties; the customer said go pound sand, and Red Hat ultimately backed down. The spook of GPL, or "no further restrictions" in particular, is strong, and Red Hat/IBM are not likely to want to test it in court.
> Analysis should involve also sources and argumentation.
It did.
> You stated a contentious belief as an obvious fact
Hang on: where?
> without source and argumentation.
I disagree.
> That's more like a commentary.
Well, if you feel that a different name is more apt, I have no problem with that.
> That's your analysis?
Are you paying for this? No? Then no, it's not. It's a passing comment.
> No it's not risible.
I would not have said so if I didn't think it. That a corporation which over 20Y has gone from being worth very little to being worth tens of billions of dollars solely from selling contracts should not consider contract law or license terms before making such an important decision?
To conclude that and maintain it seriously is laughable.
Amusingly, RH itself has contacted me officially, as well as several members of staff unofficially, and ex members of staff privately, to thank me for a cogent analysis and being fair.
OTOH, some developer types, both from inside and outside the company, are Very Angry with me on Twitter.
So it goes.
> you probably did not came across enough of various sources on the matter.
In your The Register article. It's good that here on HN you have clarified that this is Red Hat's position and your position. But that's the mistake - you should have stated this clearly in the article. Making an honest mistake is fine, pretending you don't see it at all is the reason for length of this conversation.
> That a corporation which over 20Y has gone from being worth very little to being worth tens of billions of dollars solely from selling contracts should not consider contract law or license terms before making such an important decision?
> To conclude that and maintain it seriously is laughable.
> Amusingly, RH itself has contacted me officially, as well as several members of staff unofficially, and ex members of staff privately, to thank me for a cogent analysis and being fair.
Nice deflection and PR work there. From outside, you seem to fit better in Red Hat's PR department than in a journal that prouds itself in "biting the hand that feeds IT".
Some things I have learned from now very nearly 30 years of writing for public consumption are:
* Most people cannot skim read
* They don't know that they can't
* As a result of the above, what people get out of articles is semi-random. Their eyes snag on key words and phrases or something and they come back and complain about stuff that was not there or which they totally misunderstood.
* That's why there are so many writing tools about "Fogg indices" and "reading levels" and so on; most people can't read but don't realise.
* That is also why video channels are now so huge.
I am personally not terribly interested in Red Hat's position. I approached the company -- I approached all the companies concerned for comment. I got no responses from most of them, and bland PR banalities when I got anything.
OFF THE RECORD company representatives are often happy to provide interesting details but on the record they can't, or they will get sued.
Everyone is terrified of getting sued. This is FACT #1 to remember: everyone is scared of lawyers getting involved.
So nobody will do anything that will get them sued.
RH would not have done this if it thought it would get sued, but it dare not SAY "we wouldn't do this in case we get sued" -- in case it gets sued.
Therefore there is nothing but bland corporate BS from all involved.
It is very naïve and short sighted not to allow for this.
> Nice deflection and PR work there.
And this is the proof that you don't understand me, or my position, or any of this at all.
I do not use any RH product. I do not recommend them or advocate them.
But more to the point:
Red Hat fired me.
I have no love for RH whatsoever. I have often been accused of anti-RH bias.
But you don't understand what I am really saying.
This move of the company's will make it more money short term, but long term, when I say, in print, REPEATEDLY, "this is good for the Linux world" and "good for other distros" that is because it will lose RH users and lose RH customers and lose RH money.
That will be good news for SUSE, for whom I worked longer than any other job I've ever had and a company whose products I actually like. It will be good for Debian and Canonical and Ubuntu, none of whom I have ever worked for.
> From outside, you seem to fit better in Red Hat's PR department
I should post the loud anguished criticism I have had from RH staffers about this coverage, but that would be a grotesque violation of confidentiality and of professional ethics. Many of them understand as little and as poorly as you do.
Go read my Twitter threads, and see the pages of attacks on me for my anti-RH bias.
The contract doesn't override or contradict the GPL. The GPL doesn't cover rights to future versions of the software, only the versions customers receive. The contract termination is only for future versions of the software, not the versions customers receive. Also, the contract termination only happens when a customer gives someone who is not a customer the benefit of a subscription, it presumably wouldn't happen in other cases of exercising GPL rights.
Neither are building from CentOS Stream - they are building from the same source that is in the final RHEL release, not a rolling release like CentOS Stream.
I thought they were using git.centos.org which used to be the RHEL sources?
Centos stream sources are what will become the next RHEL major or minor release. So if(?) RH is adding further stuff on top of Centos stream after forking those changes would not be available to non-customers?
Management doesn't directly know the pain of an engineer. This business heirarchy is akin to management being the body's "hands" while engineers are the "feet".
This easily accommodates the option of shooting one's self in the foot.
One of the FAANGs has/had exclusive deal with Oracle for Oracle Enterprise Linux, as Oracle charges a way less than Redhat does. Redhat/IBM figured out how many large enterprises use Oracle Linux; now they want to go after those enterprise customers to show more return on investment, as IBM paid $34B to acquire Redhat.
Another irritating move from Red Hat, likely driven by their unhappiness with the popularity of Alma/Rocky instead of a mass migration to RHEL they had probably hoped for.
As the source is still available, but not publicly, could it be that these alternate flavours enter in to a license agreement with Red Hat that's community funded?
It works out ok. The GPL only requires Red Hat to provide the source code to its users, not the whole world. A license requiring publication on the internet would be considered non-free by the GNU project. It would also be rejected by Debian because it would fail the Desert Island test [1].
What a massive hole in the GPL, and its wild it would be considered "non-free" when most people would actually consider it free. Another reason I do not like that license.
They can't, but they could do something like GRSecurity, saying that they would end their contract with / stop giving anybody who distributes it further updates.
All customers of Redhat should start daily sync of all source RPMs. That way, Redhat can't find out who is sharing source RPMs with Alma and Rocky. Right now, customers just sync binary RPMs.
"CentOS" tracked behind RHEL, and is considered more stable? Regardless, CentOS 7 is the last supported major version, and it's reaching EOL soon and will not be maintained. This sounds like a push to get those who use CentOS (free) to move to RHEL (paid) to stay in the long-cycle ecosystem.
"CentOS Stream" is a flavor that tracks ahead of RHEL, is less stable, and is really meant for those who develop on RHEL.
Since the original CentOS is no longer supported, you had new flavors come up (specifically Rocky Linux and Alma Linux) designed to fill the void to track behind RHEL.
However, it appears they rely on the open-source RHEL for their builds. It sounds like this is going away now, and the only open-source linux RedHat is gonna publish freely is CentOS Stream.
Please feel free to correct me if I'm wrong or missed something!