That’s the point. By sharing development efforts everyone wins. Seen another way, RedHat would be very different or not exist if they had to start from scratch by bootstrapping their own compiler as the first step.

Writing your own compiler is what Mark Williams did when they made the Unix clone called Coherrent. The comparison of that to current Linux distributions shows how much more is possible with the cooperative sharing of software investments between companies and interested parties.

It doesn’t take an MBA to realize how this works. If fact, it’s possible having an MBA makes the financial advantages of sharing source code more difficult to understand.

The “free to pay” issue is more nuanced than the article suggests, since Red Hat itself free-rides on the developers of much of the software it offers.

You’ll recall the issues with OpenSSL, which were basically because not a dollar of the millions being made by Red Hat flowed through to the people who wrote that fundamental software.

The Linux kernel is vital to Linux: Red Hat is a dominant and important contributor. But even so Red Hat’s contributions are about 5% (by changesets or LOC). The hyperscalers are more than that, as are the CPU vendors.

Self-support for the hyperscalers is essental. When Facebook is down it is not going to log into the Red Hat website and log a P1 case. The opportunity cost of lost advertising impressions pay for Facebook self-supporting the Linux it uses. Red Hat reducing its support rates to even 1% of its existing charge would not change that decision.

Legally, the Linux experience is dominated by the harrowing experience with SCO. If Google were to pay Canonical, then that contract opens Google to litigation from Canonical. Google makes so much money it could be worth a venture fund buying Canonical just to access that possibility of litigation with Google. Better for Google to take a free license with no support, but no contract. Far better again for Google to require its staff to use Debian, which can’t be purchased and used as a litigation vehicle.

The basic complaint of your article is that it’s unfair that Rocky Linux free-rides on Red Hat and other developer’s work. But it is somehow simultaneously fair that Red Hat free-ride on the work of other developers (who in turn free-ride on the work of other developers — as an author of networking software I am free-riding on the work of the GUI developers). The point of free software is that everyone free-rides. That’s the essential feature, not a bug.

Red Hat also free-rides on CentOS, especially in documentation. Search for an activity you might do on RHEL. Append that search with “RHEL”, compare with the results for “CentOS” or “Fedora”. Now maybe Red Hat will get around to writing the massive manuals which used to accompany Unixes from Sun or DEC, but in the meantime they are free-riding off CentOS users.

As for *everyone* paying a little for Linux — economically we call that “taxation”. Let’s take the Linux kernel: at best 15% of contributors are individuals. So you’re going to tax every Linux user (literally billions of people) and give the money to (in order by changed LOC): Intel, Huawei, Google, NXP, Red Hat, Code Aurora, Linaro, Facebook, BayLibre, AMD, IBM, MediaTek, Arm, TI, SuSE, Oracle, Nvidia. What strikes me about that list of recent major Linux contributors is that they are some of the wealthiest corporations on the planet. The very corporations we should be receiving more taxation revenue *from*.

You’ll recall the issues with OpenSSL, which were basically because not a dollar of the millions being made by Red Hat flowed through to the people who wrote that fundamental software.

The Linux kernel is vital to Linux: Red Hat is a dominant and important contributor. But even so Red Hat’s contributions are about 5% (by changesets or LOC). The hyperscalers are more than that, as are the CPU vendors.

Self-support for the hyperscalers is essential. When Facebook is down it is not going to log into the Red Hat website and log a P1 case. The opportunity cost of lost advertising impressions pay for Facebook self-supporting the Linux it uses. Red Hat reducing its support rates to even 1% of its existing charge would not change that decision.

Legally, the Linux experience is dominated by the harrowing experience with SCO. If Google were to pay Canonical, then that contract opens Google to litigation from Canonical. Google makes so much money it could be worth a venture fund buying Canonical just to access that possibility of litigation with Google. Better for Google to take a free license with no support, but no contract. Far better again for Google to require its staff to use Debian, which can’t be purchased and used as a litigation vehicle.

The basic complaint of your article is that it’s unfair that Rocky Linux free-rides on Red Hat and other developer’s work. But it is somehow simultaneously fair that Red Hat free-ride on the work of other developers (who in turn free-ride on the work of other developers — as an author of networking software I am free-riding on the work of the GUI developers). The point of free software is that everyone free-rides. That’s the essential feature, not a bug.

Red Hat also free-rides on CentOS, especially in documentation. Search for an activity you might do on RHEL. Append that search with “RHEL”, compare with the results for “CentOS” or “Fedora”. Now maybe Red Hat will get around to writing the massive manuals which used to accompany Unixes from Sun or DEC, but in the meantime they are free-riding off CentOS users.

As for *everyone* paying a little for Linux — economically we call that “taxation”. Let’s take the Linux kernel: at best 15% of contributors are individuals. So you’re going to tax every Linux user (billions of people who already paid for phones and modems) and give the money to (in order by changed LOC): Intel, Huawei, Google, NXP, Red Hat, Code Aurora, Linaro, Facebook, BayLibre, AMD, IBM, MediaTek, Arm, TI, SuSE, Oracle, Nvidia. What strikes me about that list of recent major Linux contributors is that they are some of the wealthiest corporations on the planet. The very corporations we should be receiving more taxation revenue *from*.

Now you are talking. I like the Green Bay Packers model myself–the town owns the team, the town pays, and the town benefits. I never thought of Linus Torvalds as Aaron Rogers before….now.

I understand. In the long run, Linux is what? 75 percent of the datacenter? More?

As a private user, paying for Linux in a general way, as the article proposes, is something I could not accept at the moment, as perhaps I pay something, and my favorite distro just disappears, as the developers found something else, which is more interesting, and create a new distro…
It is just too unreliable then to pay for “any” Linux or Linux in general.

Paying for RHEL? Well, if there could be any price model for private users, in a price region as in the past also Windows NT e.g. was (I was an owner right from the start with the 3.1 version; on 3.5” floppy disks), so ~350€ for 10 years support.
Problem is: Linux, the base of RHEL, isn’t owned by Red Hat, so they cannot guarantee such a timeframe, as perhaps Microsoft can with Windows, which code they own (where I would choose their LTSC (former LTSB) versions with 10 years support without feature-party; my still used Windows 8.1 internally shows as winblue.LTSB).

Linux was started as, and also stated as “free”. So, forgive me but, it is not right, if a company starts using a system as basis, which is free, will be free, and then at some point of time says “oh, we were wrong; you have to pay now in every case” (which they don’t say directly, I know). They knew what they did, when they started in the past (hopefully). They knew the licenses they have to accept.
It is correct: Every work should get a reward somehow, but in this case it is impossible to unite two sides which are that opposite: Commercial (we need to earn money) and non-Commercial (it will be and must be free).
Yes, in that way, Linux is somehow a cancer, as Ballmer stated, but a cancer that perhaps eats itself sooner or later.

(Microsoft does it very clever, from a commercial point of view: They keep their Windows, and add Linux, and now also Android, as somehow subsystems, which they adjust to their needs, with smaller effort than building both systems from scratch, as Microsoft mostly cares about “use our Azure – choose whatever system you want, but use Azure”.
Clever from a commercial point of view only, also if convenient for the users, as the problem is that they, on the Windows side, cumulate the vulnerabilities of the respective OSs by this integration.)

Perhaps their might be any other OS base in the future, which will be used under any other license, which allows to make improvements and further steps in closed-source, but then, we are again at the beginning.

Debian is altruistic, enlightened self interest, for example, that feeds Ubuntu and all derivatives – but is reliant on donors to help provide some hosting – and SPI has a non-profit stance.

Your point about the whole infrastructure being based on CentOS is an interesting one: If Facebook is using CentOS Streams, Amazon is using forked CentOS, Oracle is using forked RHEL in the same way as CentOS without paying RHEL – Red Hat have problems if they want to rejig Red Hat and rely on outside developer goodwill. This was, essentially, a huge misunderstanding on all sides but it has left a bad impression, deservedly or not.As you note, there is no such thing as a free lunch – but the people who don’t pay Red Hat provide their support in-house.

If Red Hat were to disappear tomorrow, it would leave a hole but that might be in commercial training courses, consultancy and event sponsorship as much as in the whole of the huge Linux infrastructure. The Fedora project might suffer short term but might pick up the slack or fade away: Red Hat-specific developers would find other useful work to do.As much of the valuation is goodwill as subscription, in some ways.