Back in the late 1970s, computer scientists working on the first laser printer discovered they had a problem: how do you scale a digital font without making it look terrible at low resolutions? You stop thinking in pixels.
Let's imagine the "H" of a hypothetical serif font. This "H" might have been designed to look fine at high-resolutions, but as you scale it, problems will arise because various parts of the letter such as the serifs or the stems will only partially fill a pixel. What do you with those pixels: turn them on or turn them off?
This is called the rounding problem. If you decide that any section of a letter that overlaps with a pixel should result in that pixel being filled completely, you end up with fonts that look distorted. But if you decide to leave that pixel off, you end up losing information--and that can have dire consequences, like making a vital part of the letter disappear entirely.
It was a tricky problem that at the time had only one solution: designers had to hand-tweak every font so that it looked good at a number of different sizes. Not only was this a laborious, time-consuming process, but computer storage was extremely expensive at the time. It was a waste of valuable storage to have to keep on hand 40 or 50 different hand-tweaked variations of Times New Roman on a computer or laser printer, just to display it at every resolution. You'd rather store 40 or 50 different fonts instead.
But as computer science professor David Brailsford of the University of Nottingham explains in this fascinating account of pioneering work on digital fonts, eventually, a couple of "font magicians" figured out a clever solution for the problem.
In 1982, two computer scientists left Xerox's PARC Labs and founded a new company, Adobe Systems, dedicated to finding a solution to the problem. Two years later, they released PostScript, a computer language that finally solved the rounding problem.
Here's how it worked. Earlier computer fonts stored fonts as bitmaps--basically little computer images, composed of pixels--that contributed to the rounding problem. Adobe's idea was to store fonts not as pictures, but as descriptions. Using geometric principles like points, lines, curves, and shapes, a font designer could create a font by describing what it was supposed to look like, which the PostScript system would then render into a series of pixels at the last minute, tweaking the font as necessary to make sure it rendered correctly at different sizes.
PostScript allowed designers to create just one copy of a font, and not have to worry about hand-tweaking a separate copy for every font size. And because PostScript was just a description of how a font was supposed to look, the system allowed designers to use "hints" to tell a computer or laser printer what parts of a font were the most important to render accurately at different resolutions: for example, that no matter how small you rendered an "H," the crossbar could never be eliminated.
This clever innovation didn't just allow computer printers to compete with traditional methods of printing, effectively jumpstarting the desktop publishing revolution of the mid-'80s; it also empowered designers to finally create digital fonts every bit as detailed as their print counterparts.
For more information about the early days of computer printing, check out professor Brailsford's YouTube channel here.