>>10877368
Neither

>>10877368
These days neither is really necessary unless you're making a homebrew game or a romhack or something, the limitations of games now is scope, but I imagine back in the day it was a mixture of both. You wanted to push the limits of what was possible, but there were still limits that you had to work around. Plus if you're not thinking of the game first at least to some degree, you're probably making a crappy game.

Of course, I'm not a game dev and I certainly haven't made a game for a retro platform, so what the fuck would I know?

>>10877368
Starting on better hardware like in the arcade can be a way to experiment with gameplay ideas that you would never think of on the home system, I think they never would have attempted something like R-Type or Populous on the sms or Gradius 2 on nes if they just started trying to start thinking about system limitations.
Making a game while thinking about the system limitations is valid too like in Castle of Illusion on sms. An example is wonder boy 2 vs 3.

>>10877368
Most programmers back then likely had a designer or their boss hand them a bunch of drawings and would have told them how they wanted the game to work and then it's up to the programmer to work out how to make it happen. To prevent running out of time with lots of half finished stuff you just want to focus on the most important parts first and if you have time and if the hardware will allow you can add more at the end

>>10877731
I was curious because I recall an old interview from the Atari days when they were making the Raiders of the Lost Arc game. Something about how the head creator on it was walking through the halls of the HQ cracking a whip, trying to brainstorm a means to turn the film into a video game for the VCS.

>>>/vg/

>>10877368
3. Prototype first some rough solutions, and then just polish whatever works while slapping art on top.