I work for a small company that produces a DAW and VST plugins. Supporting Linux is a huge amount of work compared to Windows and Mac. The main issue is that 'Linux' is not a thing you can support. You have to pick the distros you want to support, and then once you've picked a distro, what versions you want to support. And you need to use the C++ version that ships with that distro, so if you want to support old vers…
> And distos aren't backwards compatible. ie when libcurl4 is released, they remove libcurl3. So you can't have one binary for Ubuntu 18.04 and 16.04. If you don't want to depend on the libcurl provided by the OS, ship your own. If you don't want to depend on the glibc provided by the OS, ship your own, with your own dynamic linker. readelf -n will tell you the oldest kernel that will run your stuff. Put that in the…
Game dev: Linux users were only 0.1% of sales but 20% of crashes and tickets
331–340 of 415 posts
Re: Game dev: Linux users were only 0.1% of sales but 20% of crashes and tickets
#332Earlier quoted context omitted.
I believe nVidia is the only major GPU manufacturer left without full-featured open-source drivers. nVidia was once the best GPU manufacturer for Linux on PCs but they proved unable to make the leap from proprietary drivers to OEM-supported open source; Nouveau exists as a community effort but remains well behind the AMD/Intel equivalents. IMHO, assuming you get to choose your GPU, the only reason to mess around with…
The thing to understand about Nvidia is that they use the same model as Apple: Produce really nice software, and use it to drive hardware sales. On the driver side, a lot of the features and performance advantages of Nvidia cards are rumored to be largely implemented in software. (Think of things like having a really good optimizing compiler for shaders.) Open sourcing that code would allow other brands to copy a lot…
That's probably true, if your use case requires the absolute best GPU performance. For me, though, the Intel and AMD open source drivers have gotten good enough that the marginal improvements possible with proprietary drivers simply aren't worth the hassle. At least with AMD you are free to choose between Mesa and Catalyst, and can be assured of a decent experience either way. If you opt for nVidia hardware you're effectively locked in to their proprietary driver if you want to use even half of your GPU's capabilities.
It helps that Intel and AMD and all the non-PC Mesa drivers share common infrastructure, so improvements to one driver tend to enhance the entire ecosystem—except nVidia. As you noted, the situation is changing rapidly; I think it won't be much longer before nVidia's proprietary drivers forfeit whatever performance advantage they have left to the combined efforts of the open source community.
Re: Game dev: Linux users were only 0.1% of sales but 20% of crashes and tickets
#333Earlier quoted context omitted.
Wait what? What if only 0.5% of your costumers complain? Or that 0.1% of the pet you walk are cats and 99.9 dogs, and you get complaints that you are bad at walking with cats and bad at walking with dogs. Is it good use of the metric to conclude that the cats are the problem?
It's a good metric to conclude that you should stop walking cats, no matter what the underlying issue is.
Re: Game dev: Linux users were only 0.1% of sales but 20% of crashes and tickets
#334> We started working on the game in 2012 > Issues specific to Linux were almost entirely graphics driver related, and unique to the platform. So this is essentially a rant about linux graphics drivers circa 2012?
Re: Game dev: Linux users were only 0.1% of sales but 20% of crashes and tickets
#335Earlier quoted context omitted.
Wait what? What if only 0.5% of your costumers complain? Or that 0.1% of the pet you walk are cats and 99.9 dogs, and you get complaints that you are bad at walking with cats and bad at walking with dogs. Is it good use of the metric to conclude that the cats are the problem?
The conclusion here is stated in the article as "Issues specific to Linux were almost entirely graphics driver related, and unique to the platform." So somewhat similar to having 0.5% of your pet walking customers complaining about issues you can't do much about, or have to do extra work to make them happy - their doghouse at home is too small, their car has an issue so they can't come and pick up the cat - basically…
But I mean, without numbers on how much extra dev time the linux port demanded it's hard to reason about what a 0.1% revenue increase means. 0.1% is still 0.1% and could have been more if it didn't crash at that rate and discouraged buyers.
Since they released for Mac the game must have been programmed with cross platform functionality in mind.
Re: Game dev: Linux users were only 0.1% of sales but 20% of crashes and tickets
#336Developing software for Linux is like painting a moving train.
SDL is binary compatible on its major release (2.0) and will probably be forever at this point. libc++ only broke its ABI once when C++11 required it. What else do you need to make games for Linux? If you are using Unity or Unreal its going to statically link all its dependencies anyway. Don't use libraries that break semver.
Re: Game dev: Linux users were only 0.1% of sales but 20% of crashes and tickets
#337Earlier quoted context omitted.
Invisible demand is undistinguishable from no demand, as far as B2C is concerned.
Not true. Invisible demand is a potential market and a bunch of prospects in waiting. However that niche is way too small to be worth it.
Re: Game dev: Linux users were only 0.1% of sales but 20% of crashes and tickets
#338Re: Game dev: Linux users were only 0.1% of sales but 20% of crashes and tickets
#339Linux is the perpetual scapegoat. They selected a middleware, Coherent UI, that didn't work properly on anything but Windows. They also didn't make proper use of the Steam runtime, a mistake that continues to cause issues. Most games don't make these mistakes, so this isn't really representative of the larger state of Linux gaming. It's worth pointing out that the devs did make a legitimately good attempt at making L…
Technically speaking I'm sure you're right. In practice it doesn't really matter though, if they did it that way it's either because it was easier or because that's the way they were used to doing it. The fact that "technically" it's not Linux's fault doesn't really matter, unless you're more interested in the moral concept of guilt rather that the practicalities of making a Linux port of a game. Why should they chan…
Re: Game dev: Linux users were only 0.1% of sales but 20% of crashes and tickets
#340Linux is the perpetual scapegoat. They selected a middleware, Coherent UI, that didn't work properly on anything but Windows. They also didn't make proper use of the Steam runtime, a mistake that continues to cause issues. Most games don't make these mistakes, so this isn't really representative of the larger state of Linux gaming. It's worth pointing out that the devs did make a legitimately good attempt at making L…
Technically speaking I'm sure you're right. In practice it doesn't really matter though, if they did it that way it's either because it was easier or because that's the way they were used to doing it. The fact that "technically" it's not Linux's fault doesn't really matter, unless you're more interested in the moral concept of guilt rather that the practicalities of making a Linux port of a game. Why should they chan…
If you fail to use their runtime properly, you're doomed to suffer from Linux fragmentation problems.