> yet multi-billon dollar studios claim supporting anything beyond x86_64 Windows is an impossible feat.
It makes sense though, there are different circumstances in play.
One day back in ~2013 (or so) as i returned from work, a gamedev studio, i bought my first Android device (some cheap tablet that was sold in a basket) from a mall i often visited after work. Once i arrived at my place and checked it out for a bit, i decided to port my previous (current at the time) 3D game engine on it - downloaded the SDKs, read some docs and tutorials, etc and banged out code until some hours later i got the test game running[0] on the device.
All it took was some spontaneous decision and a few hours of my time at home. That's about it.
However if something similar (porting to a new system) was to happen for the game engine i worked on at work, it'd take much more effort - even after assuming the decision was already made. Different programmers worked on different aspects of the engine that would need to work on it (graphics, audio and low level / system support would be the least) and as the engine relied on middleware we'd need to also ensure the middleware supported whatever we wanted to target, the programmer responsible for that middleware had to take that into account and ensure the legal side was covered too (some middleware consider ports -or basically any executable you make- as separate licenses). We'd also need to have the buildmaster work into integrating the new platform for the automated builds and testing (and perhaps write some basic tests if needed). QA would need to allocate time to test on the new platform, not only for the platform specific functionality but also some tests would need to run on the new platform even for functionality that had nothing to do with it to ensure all new stuff worked as expected (this in turn could cause other ripple effects - for example some new functionality in the engine that worked in a powerful platform might prove too demanding/slow for the new platform - at that point someone would have to decide -meaning meetings, etc- if the new functionality will remain, if it will be altered to work on the new platform -this means some research time will be spent on this- or if it will become an optional feature that somehow is only available to the powerful platforms while not being available on the new one -which would require not only the programmer time to implement this switching but depending on the functionality, potentially also artist/designer time to specify where it will be used- but also depending on what that'd be about, it may require some fallback functionality too -again more programming time).
And that would be for a small-to-midsized AAA game developer at the time with a rather small engine team - in larger teams and companies there'd be way more people and friction involved. All that would translate to a lot of extra time and thus cost.
If you are an indie developer it is very easy to just add support for something, but much harder to do if you are a multi-billion dollar studio. A couple of years after the above i joked i could port my engine (same as the Android one) to Haiku if i felt like it - and then i did that[1]. Meanwhile that wouldn't even pass as a joke in the companies i worked at.
It makes sense though, there are different circumstances in play.
One day back in ~2013 (or so) as i returned from work, a gamedev studio, i bought my first Android device (some cheap tablet that was sold in a basket) from a mall i often visited after work. Once i arrived at my place and checked it out for a bit, i decided to port my previous (current at the time) 3D game engine on it - downloaded the SDKs, read some docs and tutorials, etc and banged out code until some hours later i got the test game running[0] on the device.
All it took was some spontaneous decision and a few hours of my time at home. That's about it.
However if something similar (porting to a new system) was to happen for the game engine i worked on at work, it'd take much more effort - even after assuming the decision was already made. Different programmers worked on different aspects of the engine that would need to work on it (graphics, audio and low level / system support would be the least) and as the engine relied on middleware we'd need to also ensure the middleware supported whatever we wanted to target, the programmer responsible for that middleware had to take that into account and ensure the legal side was covered too (some middleware consider ports -or basically any executable you make- as separate licenses). We'd also need to have the buildmaster work into integrating the new platform for the automated builds and testing (and perhaps write some basic tests if needed). QA would need to allocate time to test on the new platform, not only for the platform specific functionality but also some tests would need to run on the new platform even for functionality that had nothing to do with it to ensure all new stuff worked as expected (this in turn could cause other ripple effects - for example some new functionality in the engine that worked in a powerful platform might prove too demanding/slow for the new platform - at that point someone would have to decide -meaning meetings, etc- if the new functionality will remain, if it will be altered to work on the new platform -this means some research time will be spent on this- or if it will become an optional feature that somehow is only available to the powerful platforms while not being available on the new one -which would require not only the programmer time to implement this switching but depending on the functionality, potentially also artist/designer time to specify where it will be used- but also depending on what that'd be about, it may require some fallback functionality too -again more programming time).
And that would be for a small-to-midsized AAA game developer at the time with a rather small engine team - in larger teams and companies there'd be way more people and friction involved. All that would translate to a lot of extra time and thus cost.
If you are an indie developer it is very easy to just add support for something, but much harder to do if you are a multi-billion dollar studio. A couple of years after the above i joked i could port my engine (same as the Android one) to Haiku if i felt like it - and then i did that[1]. Meanwhile that wouldn't even pass as a joke in the companies i worked at.
[0] https://i.imgur.com/9P3C2zj.png
[1] https://i.imgur.com/cI97myO.png