Hacker News Reader

Comment by someexgamedev | Hacker News Reader

Middleware company is right. Supporting the platform is not worth the trouble. Data shows again and again that Linux gamers will be a fraction of a percent of your customers but generate an outsized amount of support tickets. Most of these support tickets will require holding their hand through basic Linux box administration.

View on HN

> Most of these support tickets will require holding their hand through basic Linux box administration.

Yeah, customers can be so dumb.

https://github.com/valvesoftware/steam-for-linux/issues/3671

rm -rf "$STEAMROOT/"*

Now, don't get me wrong. I have a dedicated win10 partition for gaming. I gave up that fight a long time ago.

I game mainly on Linux (since most games interesting to me happen to be on Linux).

Valve has been and is doing a lot of work to expand Linux support. I wonder where Linux gaming would be today without them.

I thought for sure SteamOS and the various Steam Machine “consoles” would create a critical mass, but it seems to just be going considerably slower than I would have predicted. I’m hopeful, but I don’t see the economics adding up right now

Critical mass starts with you. If you want to play video games on Linux, support developers and games that you can plan on Linux.

You're not helping your case by pointing out the platform's standard practice for installers is to use a language from the '70s that makes it trivial to write "rm -rf /", and (as that ticket indicates) the command will carefully work its way through all connected devices, including external drives.

You are not helping your case either by pointing out a dev's inability to handle a well documented language with known pitfalls that has been in use since the 70's.

If they can't write bash scripts there are a lot of options to install, update and remove software with built-in fail-safes.

That language also makes it trivial to make a safe wrapper function for "rm" which validates the arguments and prevents accidents.

Developers should be quite right then ditching that middleware company and finding something professional instead, that works on Linux.

Most common closed gaming middleware works on Linux today (Wwise, Umbra3D and so on). And that's besides open source options of course. So the above is very unprofessional view.

Wwise only works on ALSA, it doesn't work on OSS or JACK or ESD or whatever some user's preferred sound daemon is. Or at least didn't when we had users complain about it. Fragmentation is a serious issue, and Linux users demand that developers implement one solution because they believe it to be technically superior.

Doesn't it work with PulseAudio? Using raw ALSA doesn't sound common. And there are plugins for such cases.

Linux as a whole is moving towards Pipewire anyway: https://pipewire.org

That will take time still.

Now yet ANOTHER way to do audio on Linux? Oi vey...

https://www.xkcd.com/927/

It's a good evolution over Pulse and Jack. Partially driven by Wayland security model too. So I see no problem in doing the right thing here.

And I don't like that xkcd in general because it's missing the point. Its basic idea boils down to "don't improve anything because it's another new thing" which is outright wrong. Not every design can be improved incrementally.

It's still possible for a new interface to provide a compatibility or translation API for prior software.

Sure, and it's already done for Pulse / ALSA for example. Pipewire can do the same if it's possible.

Pipewire can also route video, silencing once and for all the "muh network transparency" whiners who are dragging their feet on switching to Wayland.

You can use ALSA apps on pulseaudio systems. Pulseaudio and ALSA covers really big percentage of Linux installs.

I've always felt that choosing Linux implies some sort of technical proficiency, or at least willingness to learn. Are those days over?

Linux implies a level of user arrogance and heavily implied it's the developer's fault and inability to understand how things the way they are. We can only test on so many configurations, and users will complain when their distro, or sound server, or preferred desktop environment isn't supported. If you don't think that last one matters, what happens when a certain obscure floating WM doesn't focus the window in an ICCCM-compliant way (WM_TAKE_FOCUS) and as a result, no keystrokes are delivered to the window. I had to fix that, for one user who was very loud on Twitter about how we clearly don't care about Linux. The total number of Linux sales was less than 0.1%, but cost us way too many support tickets and weeks of developer time.

That's pretty silly. If you have enough technical knowledge to use a non-default window manager, I feel like you should have enough self-awareness to know that you shouldn't necessarily be catered to by devs. I primarily use Linux on servers, though. Maybe I don't get the mentality of a Linux Desktop user.

Sounds silly, yes.

OTOH I've never met one of those users, so unless more devs chime in to tell me this is fact ir someone can point to a number of mailing list post, issues or other verifiable artifacts I'll be tempted to assume it is a strawman.

Sounds more like an anecdote than a strawman.

Linux on the desktop is very nice and easy to use, once installed & configured.

However most people dont know how to install or configure Linux, Windows (or any other OS) on their devices.

This also happens to be the reason why Linux is not popular on the desktop; it lacks the strong relationship with the OEMs that gets an operating system pre-installed.

Linux works best for people who just install it and use it, -and for people who are willing to learn how to fix it if they break it, AKA "grandparents and/or techies".

About 10 years has passed since I concluded that Ubuntu was more userfriendly than Windows. I had "installed"[0] Windows on someones computer and it took 4 hours to get it finished from pressing the power button the first time until bthe last nagware antivirus was gone and the computer was ready for use[1].

By comparison an Ubuntu installation at the time was 20-45 IIRC minutes depending on disk speed.

Ubuntu would the be ready to write a document, create a spreadsheet and browse the Internet, everything the typical user would wish.

There's only a few things Windows would do better at that time, most notably running Windows-only software and look familiar to Windows-only users.

[0]: or configured, it was supposed to be preinstalled so no idea why it used two hours just to get me to a login screen.

[1]: I've been a sysadmin so I'm well aware that Windows can be slipstreamed etc, but my experience here is what an end user would see.

I've seen this sentiment quite a bit; it's an oddly romantic view of systems administration.

Most of the time you're hunting through documentation to find the correct settings. It's nothing new in general, it's not some kind of exciting math that's new to you, and you're rarely actually learning how the thing works. It's mostly trying a series of things until you figure out that some strings go in a place.

I appreciate it more when I'm writing software, but largely because I can actually invest the time to really solve the problem, that it fucking stays fixed, and that people will get back hours of their life that would otherwise be spent getting frustrated and searching aimlessly.

Reply on news.ycombinator.com