is Unity a good engine for retro/old school open world game?

yesterday a thought came to my head that most modern games don't come to their full potential because of the CPU and GPU restrictions, what I mean is that if the current household technology would be more advanced, games will be more advanced, the CPU and GPU in my opinion are the bottle neck for modern games (each game aim to run at 60 fps for the best graphics that is available by the best computer you can buy today).

so... I was thinking that if games like GTA would have a better systems to run on you could see not 1 or 3 cities but 10, maybe even a whole planet (so if I'm right about this, not only graphics influenced by the technology each household hold but also gameplay, sizes, advanced AI, etc).

then I thought (as a gamer who prefers gameplay over graphics) why not make a low graphic game, but make everything else better.(which I personally don't understand why no serious studio ever do, do all the gamers just want better graphics? I would love if someone could en-light me about this.)
when I say everything else better, I mean EVERYTHING, I would love to see a game with the old Duke Nukem 3D style graphics (eg. instead of models for enemys and guns - they made of planes with animated sprites), but in a open world, which every city scaled just like a real city, AI system that handles dozens of people.

so basically my questions are, is Unity a good engine to make something like this? will UDK be better for this kind of project? would you like to play this game?

 

The first thing people's ego's `see` is usually the appearance of something on the surface, the first-glance first impression, so that means it's important for these companies to make a good first impression with lots of flash and pazazz and high-quality-looking content... that may mean that the game actually isn't that great or fun to play once you get into it, the actual experience of it might be dull, but it's the flash and pazazz that might get you to be impressed enough up-front to be more likely to buy it. It's a `make more money` thing, not necessarily a `make a better game` thing. Making a better game may therefore not necessarily lead to the maximum amount of money.. but I'm sure there are exceptions.

I think as has been said elsewhere you need good graphics + good gameplay executed well and in a way that is entertaining and interesting... doesn't necessarily have to be realism-styled `serious` graphics. More symbolic, pixellated, retro, colorful, or whatever graphics can be fun too, sometimes moreso.

 

I've been working on a pixel world and music for my unity game, I've never attempted an artstyle like that in udk but I'd assume it to be pointless. You would be better ofc using lwjgl or cocos for a game like that. I would avoid UDK anyway due to its royalty charges.

 

Voxel open world game? That would be excellent!

 

While simplifying the graphics will definitely buy you some room to expand the scope, don't underestimate how much processing the pure simulation can consume. Dwarf Fortress, for instance, is an extreme example of what you're talking about (ASCII art for graphics, insane detail of simulation) and it's actually surprisingly processor intensive. Ditto Minecraft, although that does have (fairly simple) graphics.

Part of the reason is that while 3D graphics add a lot of processing, the GPU is specialized to do that work, and will probably be just sitting there otherwise (although it is possible to make use of it). I would say the largest gain in using simpler/less graphics is the time saved in development - it's a hell of lot less man-hours adding a background and some text than a full environment, characters, and voiced dialog.