The Pyro wrote on Mar 25, 2016, 09:28:
Honestly, the last thing Carmack did that was great was the Quake 3 engine.
I think he deserves some credit for what he tried with the Doom 3 engine. It was one of the first games to really take a stab at universal per-pixel lighting. The game definitely had some great moments - I still love all the shadows and lighting effects from imps hurling fireballs in dark rooms. Unfortunately the hardware of the time couldn't quite handle Doom 3's demands, and id was forced to sacrifice quite a lot of other things (e.g. texture resolution) for performance reasons.
This is why I wouldn't give him credit for the Doom 3 engine. It was ahead of its time, and that was a bad thing. He focused on lighting and shadows, as you mentioned, but hardware wasn't ready for it. As a result, quite a bit had to be sacrificed. Doom 3 gameplay was designed very specifically around the lighting and shadows - whether this was because it was what the engine did best or simply how things developed I do not know, but that was a bad thing, as well.
Ultimately, it was an engine really good at one thing, but nothing could handle that one thing, and not particularly good at other things. As a result of this (and the tools not being as good as the competition), Prey is the only game made with the engine not directly connected to id itself.
It was a bad game engine for the time. It was a bad business decision. It resulted in a bad Doom game. From a technical perspective it was impressive, but it didn't really work. This, to me, makes it a failure. John's job was to create powerful engines that make impressive games. id tech 4 was a tech demo.