A key advantage of PostScript is that you can scale all output using transformations, so that you internally work with the same coordinate space (called user space) on all devices, and implicitly scale — and also for example rotate — all graphics when they are actually drawn.
This extends to fonts too, since PostScript letters are drawn using the same graphic primitives as other graphs.
I worked on a project in the early 1990s where we used the utterly wonderful HyperNeWS for most things - we used to play tricks on each other by doing silly things like logging into someone else's NeWS server using psh and setting the rotation in the transforms for a terminal window to be slightly off as people were using them - eventually people would notice of course but things would just work.
Another neat trick that HyperNeWS supported was drawing a shape in its graphical editor and then pasting that shape as the shape of a window...
My friend and I had an ongoing competition / psychological experiment to psh into each other's NeWS servers and see how long it was before we would notice a some semi-subliminal or mildly psychedelic window system hack.
Circa 1990 I was working/studying in university on a mix of 1024x1024 1 bit per pixel XTerminals and workstations (Sun, Apollo?) with somewhat better resolutions and colors. We had one NeXT too. Printers where mostly dot matrix, with some lasers, A4 format (Europe here). Nothing would have prevented sending a Postscript screen output to a printer, with some white space around due to different geometries. Actually I learned Postscript in 1996 to print some stuff directly instead of drawing it using Word or whatever.
This extends to fonts too, since PostScript letters are drawn using the same graphic primitives as other graphs.