CIG Chad McKinney@CIG_ChadMcKinney
I see this kind of response from time to time so I thought I'd give some perspective on why this approach may not make the most sense for a very large endeavor like Star Citizen. First is the issue of finding someone that will license something to you that is actually what you need. There's of course lots of big games that have been made, and many engines a long the way. The industry has indeed re-solved many problems in doing so, for example the recent Unreal 5 tech demo showed off some crazy graphics, which is likely an evolution of their voxel cone tracing implementation that was in early access in UE4. What you may not know is that there's already other implementations of the same approach in other engines, even Ogre3D has one right now (
https://www.ogre3d.org/2019/08/05/voxel-cone-tracing). Why didn't Epic just use one of the existing ones or vice-versa? Well for one, licensing fees or license issues, but on a technical level because another implementation is built in a very completely different context. Often times you'll spend just as much time trying to integrate someone else's solution as just writing it yourself, but when you write it yourself you can make it very specifically solve your problems in the most direct way in the context your engine/game exists in without any bloat of a generic implementation. Now there's always a balance you strike, and you'll still find a fare amount of middleware in our engine, such as Wwise, which we haven't replaced or attempted to rewrite. It's about picking your battles, and I can tell you that it is often a mistake to bet highly on your most important fundamental systems on some other entity and hope that their solution really fits your needs and you will get the support you need. That doesn't even get into where you take your tech from there, or if you plan on making multiple games or may want to license to the engine yourself. Video games and other high performance programming domains benefit greatly from very highly tuned specific solutions which is why companies in 2020 many very large companies are still developing their own engines and haven't just converged on a single industry solution for an engine.
To use your example of Elder Scrolls Online, their problems are very different from Star Citizens. The inventory system is much simpler, and their world persistence is non-existent. ESO highly utilizes sharding with dynamic instancing, wheras in SC we are trying to push towards a unified singular instance in any region and possibly globally. It wouldn't even begin to make sense to try and take their infrastructure and integrate it into ours to solve the problems we want to solve. Source: I worked on Elder Scrolls Online.