The Tragedy of Systemd | Hacker News Reader

The Tragedy of Systemd

youtube.com

21 pointsjcamou11 comments

This is a click-bait title.

This 32 minute talk is about:

   what came before systemd (e.g. init)?

   what is systemd trying to solve?

   Lennart Poettering is misunderstood and shouldn't
   be getting death threats

   what can systemd teach FreeBSD people?

   how can systemd inspire improvements to FreeBSD?

A really excellent, thought-provoking talk. When the speaker declared UNIX to be essentially dead, it would have made sense to stop and take questions. That's the conclusion! Linux won the Unix wars. Debating the merits of systemd and whether to adopt it in BSD is irrelevant. BSD is in decline and will slowly fade away.

Linux (the kernel) is arguably in decline as well, especially now that it's clear Android and ChromeOS will both pivot to Fuschia/Zircon.

Linux has almost 3 decades of development and experience behind it, will take a lot of time and money for Google to catch up with that.

Linux has multiple communities and companies behind it, it won't disappear overnight if one of them go bust.

Linux can run on anything, Fuschia will most likely only run on things that Google cares about.

Most server software today runs on either Linux or Windows Server.

What makes you think the Linux kernel is in decline?

OSX has decimated the Linux on consumer baremetal use case. It's increasingly rare to meet anyone, even those who interact with Linux systems daily via the cloud, who has actually installed and run it on their own hardware let alone use it daily that way for significant amount of time.

I can't predict how long it will take Google to get Fuschia in good enough shape for Android/ChromeOS production use, but the intent is clear and confirmed.

The cloud is the only place where the kernel still sees a lot of usage, but containers and platforms like Kubernetes serve in part to weaken Linux's position in that space. WSL for example will very likely be utilized for running Linux userspace containers natively in Azure, and knowing Microsoft that will immediately be followed by Azure-specific features being made available through proprietary extensions making the next versions of those containers some combination of a Linux-compatible userspace mixed with WSL-specific extensions for the Azure-specific magic.

I've been heavily involved in Linux since 90s, and it pretty clearly looks to be in decline from my perspective. Hopefully puri.sm can get their mobile device off the ground successfully and get to a point where it's a sustainable business. It would be pretty sad if in another decade the only time we encounter Linux running on baremetal is on gadgets for tinkering like the raspberry pi.

> OSX has decimated the Linux on consumer baremetal use case.

By consumer baremetal you probably mean laptops and desktop PCs, because if I would include routers, NASes and phones, your argument would be dead on arrival. Even if we limit ourselves to laptop/desktops, OSX did not decimate Linux, because the Linux desktop never actually had a high market share to begin with. All the charts I have seen show that the Linux Desktop market share was consistently low, the existence of the Mac didn't affect it much.

The 3 main reasons the Linux desktop isn't reaching bigger audiences are:

- driver issues, which is partially because of the way kernel development works - no one good consistent UI - too many distros and versions and configurations to support

If you would erase the Mac from history, these wouldn't be affected.

The consistently low market share has little to do with the kernel.

> but the intent is clear and confirmed.

The interest was there with Google Wave, Google Reader, Dart, Google Plus, Picasa and Google Desktop Search as well among other things, companies can change their mind.

> but containers and platforms like Kubernetes serve in part to weaken Linux's position in that space.

Containers were made possible by Linux kernel features, the container images are all Linux, I don't see how that weakens it, at the end of the day, your application still runs on Linux.

> followed by Azure-specific features being made available through proprietary extensions making the next versions of those containers some combination of a Linux-compatible userspace mixed with WSL-specific extensions for the Azure-specific magic.

I'm failing to see what does this speculation has to do with the original argument.

> It would be pretty sad if in another decade the only time we encounter Linux running on baremetal is on gadgets for tinkering like the raspberry pi.

There are plenty of laptops and desktops that work well with Linux if that's what you need, there are vendors like System76 that are dedicated to giving you just that.

I think one of the reasons why people like me use Linux is to avoid big tech companies.

I will avoid Fuschia specifically because it was made by Google.

Do you use Chrome? Android? iOS?

I've been using GNU/Linux since 1993.

It's turned into a bloated mess, complete with a Foundation that is basically run by Microsoft.

It is overrun by idiot noobs doing unnecessary, unjustifiable things.

NSA sneaking backdoor encryption into the kernel? W.T.F.? How does that even happen.

That BSD family is looking better and better, let me tell you.

Yeah, except now we have Linux distribution wars instead.

A really great talk on the subject.

Reply on news.ycombinator.com