Hi josh, in your opinion what was the best game engine you and your teams worked on that were the best to work with and what was the worst one when working on development of games.
I have a hunch that most experienced game developers think about engines less in terms of “best” and more in terms of “least bad” or “best for this type of game”.
IMO, Unity is the best engine for small scale games and small teams. You might be thinking, “Like Pillars of Eternity?” No, in retrospect I actually think both Pillars and Deadfire are too big for Unity. But if you have a team of 1-10 people, it’s fantastic. It’s easy to rapidly prototype in Unity and the editor is easy to work in.
Unreal is pretty good for larger scale games. I don’t think it’s as great for small teams because there’s more overhead just to getting things up and running (again, IMO). I have to also qualify my opinions on Unreal because I’ve never directed a game using it, I’ve only worked on it as a staff level developer.
Obsidian’s Onyx engine started out being pretty nightmarish (Aliens-era) and seemed to get a lot better on subsequent projects (Dungeon Siege III, South Park: TSoT). The area pipeline was incredibly fragile/hard to work with and scripting was practically nonexistent, a reactionary consequence of programming leads having bad experiences with our over-scripted games like NWN2 (see below).
If you wanted to do more or less exactly what it did, the Infinity Engine was pretty easy to work with. That said, there was a boatload of broken and bad stuff in it / the editor when we were working with it. But in terms of making content, the pipelines were pretty straightforward. All of the scripting was LUA and even though I didn’t do much of it, was also pretty straightforward.
I came onto NWN2 fairly late, but the NWN1 engine and editor were actually pretty easy to work with. By replacing the editor and renderer (things that were arguably technically necessary) we made it a lot harder to work with, even if it was more flexible. Neverwinter Script is very powerful, powerful enough for junior designers to write scripts that turn into nooses for them.
Bethesda’s engine was the easiest to create content for, by far. Source control was easy, iterations were fast, the scripting language was pretty powerful – just easy to work in. Not necessarily easy to
change, but if you wanted to do what we did on F:NV, which was make a bunch of new content and new features for the F3 engine, it was great.
The RenderWare engine was not awful to work with (at Midway), but RenderWare Studio was very slow and painful. I did not have a good experience working in that editor.
When Black Isle was using LithTech (1.0) for Torn, I would argue that the engine was not even ready to be used as middleware. One of many problems that plagued the Torn team.
The Jefferson/Van Buren engine was kind of a pain to work with on Jefferson but became much easier to work with on Van Buren. Still, the area pipeline was slow and fragile because of how lightmaps were baked/exported.
I think that covers all of the engines I feel comfortable opining on.