I think the true generational divide was between DOS and Windows. After spending an inordinate amount of time in DOS (making the journey to Borland C), it felt awesome that you could draw onto the screen, no questions asked. I made so many games and graphical programs just drawing to the screen and then reading the input in a loop, it worked great.

Then I transitioned to Windows, and had to contend with all the complexity.. What's a Window Class and how do you write an event loop to respond to window messages, why is this all so complex.. Why is MFC somehow not really better.

Having written a simple DOS GUI (that kinda looked like Windows), I had to discover that a Windows app of equivalent functionality took more and messier code.

And forget drawing to the screen, you had to grapple with GDI drawing contexts, it was slow and bad.. (Not like Linux was any better than this regard).

As for DOS, I have had the privilege of hacking together a bitbanging VGA adapter, so I know for a fact that drawing to the screen is really as simple as setting values in an array, then, scanning it out to the display.