Red Hat vs. SUSE vs. Canonical Contributions To The Mainline Linux Kernel Over The 2010s

[bregma]

How valid is the assumption that contributors always use a corporate email address?

I know I've made a lot of contributions to open source projects, but since I change employers I just do it under my own domain. Few people get jobs contributing to the Linux kernel without previous experience. It's not terribly convenient to re-obtain contribution credentials every time you change jobs. It also makes you a target of enmity from people like our gentle readers here on Phoronix if they see you have an email domain that tickles their amygdala in an undesirable way. My guess is most active and successful kernel contributors do not, in fact, use one of the corporate emails to make their contributions, so analyses like these are really just worthless clickbait fluff.

Still, if this argument applies to all employers, then it would be far more meaningful to see a per-capita breakdown of the numbers.

[ALRBP]

Be interesting to see stats on what part of the kernel these companies are mostly touching.

Could you do these graphs other parts of the system too? GNOME, systemd, mesa, dbus etc.

Why GNOME ? (and not KDE)
GNOME is not at all a fundamental part of a GNU/Linux system. AFAIK, GNOME is used by a minority of desktop systems, and not at all by server systems.

[birdie]

Michael, these comparisons are not only worthless, but they are also simply bad because people will take them at face value without thinking twice.

These companies are related to Open Source quite differently which means they will have quite different amounts of contributions. For instance, RedHat employs core GTK/Gnome/kernel/SeLinux/SystemD/PulseAudio/etc. developers, so it becomes immediately obvious Suse and Canonical will trail RH by a large amount.

We've already seen how much people hate NVIDIA and in your previous piece where you compared NVIDIA's contributions to the kernel vs Intel's and AMD's which proves that even further despite the fact that the latter two companies produce a whole lot more hardware that needs support from the kernel.

Please stop. Maybe it's a slow news day but we'd better off without having the things which could only instigate hatred and animosity.

[Britoid]

Why GNOME ? (and not KDE)
GNOME is not at all a fundamental part of a GNU/Linux system. AFAIK, GNOME is used by a minority of desktop systems, and not at all by server systems.

Can do KDE too! Was just using it as an example.

I know GNOME has people from multiple companies working on it.

[Mez']

Red Hat employees: 13,400 (https://en.wikipedia.org/wiki/Red_Hat)
Suse employees: 1750 (https://en.wikipedia.org/wiki/SUSE)
Canonical employees: ~650 (I work here, but Wikipedia says 443).

Obviously, I'm biased but might be nice to do some of these numbers per employee as well.

This is not a bias. You can make some people believe anything you want out of gross statistics (even that Romelu Lukaku is a good football player). They don't mean much.

What would be the most interesting is to compare these relatively to Turnover/Sales revenues. Ultimately, that would indicate how much each company dedicates of their forces to contribute.

[Britoid]

Upstart - dead
Unity - dead
MIR- dead (yes, I know that MIR is still developed but not in the originally intended form)
AppArmor - don't even
snap - we will see how this one will turn out but probably dead after all

Great effort.

You forgot Bazaar, Xsplash, Usplash, Launchpad, Ubuntu Software Center, Ubuntu One and the best for last, Ubuntu Edge.

[Mez']

You forgot Bazaar, Xsplash, Usplash, Launchpad, Ubuntu Software Center, Ubuntu One and the best for last, Ubuntu Edge.

Oh boy, here we go again. Aren't you ever tired?
And by you, I don't mean you in particular.

A lot of these were really good (Ubuntu One and Unity, I miss you). Please respect that others might have liked some of these and stop trolling.

[Britoid]

Oh boy, here we go again. Aren't you ever tired?
And by you, I don't mean you in particular.

A lot of these were really good (Ubuntu One and Unity, I miss you). Please respect that others might have liked some of these and stop trolling.

I didn't say they weren't good, only that they're dead.

[Mez']

I didn't say they weren't good, only that they're dead.

Sure. But why does the discussion around Canonical always go down to that? Why would some here (and elsewhere) be so angry all the time at Canonical?
Is there some kind of underlying frustration because it is undoubtedly the most popular and used Linux distro? I just don't get it.
I'm not even spending one second bashing on other distros, although I could (Red Hat made some terrible decisions regarding Gnome). But I don't see how it is constructive to always end up some topics with the same boring stuff (when it's not common places).

Again, I'm not pointing fingers at you, I'm just growing annoyed by the propaganda (it is) of some people.

[eidolon]

Article mill must produce articles to serve ads. I'm not saying this article is worthless (indeed, some readers requested it) nor has Michael used baiting language in it, but the door has been left open to the assumption of bad faith and there needs not be any.

I can understand looking at profit (outside the scope of this article) and determining which company is in the best position to hire more programmers (that would be Red Hat, by a mile), but then again, it's not about hiring any old coder (skill level is important) or adding copious redundancy if a desired function is already being accomplished successfully without undue strain on existing employees. I can also understand looking at commits per corporate domain, but then again, if Canonical fulfilled its business needs with around 745 kernel commits in 2019, SUSE fulfilled its business needs with around 2,076 kernel commits in 2019, and Red Hat fulfilled its business needs with about 3,864 commits in 2019, fuck it. This isn't a morality tale (Red Hat bad, Canonical bad, etc), it's just data. I am probably more interested in whether any of these (or any other) companies desire to scale up or scale down their level of involvement.

Red Hat employees: 13,400 (https://en.wikipedia.org/wiki/Red_Hat)
Suse employees: 1750 (https://en.wikipedia.org/wiki/SUSE)
Canonical employees: ~650 (I work here, but Wikipedia says 443).

Obviously, I'm biased but might be nice to do some of these numbers per employee as well.

For this article, we are only concerned with functionally relevant employees (i.e. Linux kernel contributors), correct? We aren't concerned with other functions like custodial services, security, office support staff, marketing, programmer contributions to projects other than the Linux kernel, etc. or the degree to which each individual company has these non-kernel-related functions fulfilled in-house (i.e. included in employee head count) or contracted out (and thus not included in employee head count).

All of these companies have staff contributing to other areas, too, so if you are going to measure by employee, you might want to look at more than just the kernel. Yes, I understand that is nearly impossible to measure.

This.

How valid is the assumption that contributors always use a corporate email address?

And this.

These companies are related to Open Source quite differently which means they will have quite different amounts of contributions.

And this too.