Hacker News new | ask | show | jobs
by wltprgm 2075 days ago
They mentioned Vulkan and OpenGL but no sign of supporting Linux, anyone knows why at all?
2 comments

Maybe my rant was a little bit too short . But to understand, one has to know that Roblox used to be working well enough on Linux through Wine.

Then, up to the current day: "Roblox anti-cheat started detecting Wine and would force the game to quit."

More details here: https://roblox.fandom.com/wiki/Roblox_on_Linux

I have been a ROBLOX player since 2009. The Linux anti-cheat began to be enforced because practically no-one used the Wine port for honest gameplay; it was almost always to do some kind of memory manipulation or for script kiddies to execute exploits.

Not many people outside of the ecosystem are aware but ROBLOX was actually really late to the concept of dividing client and server responsibilities. It used to be that one client could actually propagate changes out to all the others; therefore if you could manipulate memory you could arbitrarily execute code (within the sandbox) on all other players. There used to be popular scripts like Person299's admin commands that could be injected into any server and allow the exploiter the ability to run chat commands like "kill/<user>" or "ban/<user>" (self explanatory). Only a few years ago did they introduce RemoteEvents and ServerScriptService which finally allowed developers to ensure some code could only execute server side, and properties like FilteringEnabled which blocked clients from propagating local game changes out to the server and everyone else. This broke hundreds of thousands of old games, which was probably why they held off, but nowadays any new games are using these new features to prevent easy exploitation like in the past.

My guess is they have nothing against Linux now, because they have implemented proper security, but the only thing stopping them are the technical challenges, and there are not enough people in their target demographic who use Linux to make porting worthwhile.

A common trend I see is that game developers tend to target one platform after another instead of developing cross-platform from the start. This lowers the initial cost, but increases the additional cost for each platform as less of the previous implementation can be reused. This leads to thinks like the discarding Vulkan because it's somewhat harder than Metal for the macOS port and then having to still do the Vulkan work for Linux which may cost too much for the expected Linux sales even thoug the the difference between Metal and Vulkan might have easily been covered by the Linux income.
Traditionally in the industry many studios focus on a specific platform and there are agencies whose whole business is to do ports.

GNU/Linux just doesn't provide revenue to make it worthwhile.

https://store.steampowered.com/hwsurvey/Steam-Hardware-Softw...

Now that is a bleeding generalization. As long as Feral is in the Linux market we know that Linux ports can be profitable. There have also been developers that have provided sales numbers for their Linux ports and for many of them porting has been a success.

Maybe that won't be the case for all games, but low market share does not tell you much. Remember that the platform-specific cost - even with rewriting the renderer - of a game will in most cases be dwarfed by design, voice acting, localisation, marketing, etc which all are already covered by the existing target platforms.

Your assertion is also not well defined since "worthwile" could mean profitable (which will be the case for third-party porting companies) or more profitable than other development efforts.