To develop a new computer font, you can simply tweak an existing print font. But it’s not usually a good idea. Typefaces that look beautiful on the page often look terrible onscreen. Especially at smaller point sizes, rounded edges and diagonals look ragged, spaces within and between letters close up, and fine lines disappear.

The better approach is to start from scratch. An early example is the Lucida family of fonts, created by Charles Bigelow and Kris Holmes, who paid particular attention to the fonts’ legibility, especially at low resolution. Verdana, released in 1996 and designed by Matthew Carter, was the first typeface that Microsoft created just for use on computers. Also designed to be readable at small sizes, it has many features to enhance legibility on screen: lowercase letters that are proportionally tall compared to uppercase letters, stroke widths that aren’t too thin, and generous spacing both inside the letter and between letters. Well over 90 ­percent of Windows and Macintosh computers now have Verdana installed on them, making it one of the most widely available typefaces in the computer world.

Microsoft’s typography group wanted to include several new screen-friendly typefaces with Windows Vista, so in 2004 it staged a competition, inviting some of the world’s top type ­designers to enter. Of the 26 submissions, six Western fonts were selected, and Microsoft then hired each winning designer to design the entire typeface. The results are two serif faces, called Cambria and Constantia; two sans-serif faces, Calibri and Corbel; a flared-serif face, Candara; and a monospaced face for programmers, Consolas. These six fonts are now shipping with the new operating system.

A good example of how the new fonts were optimized for onscreen viewing can be seen in the lowercase letter “g” [see figure, “What’s in a Letter”]. In a typical “g,” the top edge of the lower arc, or bowl, angles slightly downward. But in the Vista fonts, each lowercase “g” has a straight horizontal bar across the top of the lower bowl, so the letter appears crisp.

The new Japanese font that’s included with Vista is in many respects even more impressive. Japanese kanji characters—there are tens of thousands of them—tend to contain more strokes per character than do Western letters. So a particularly complicated character might have more horizontal lines than there are pixels to represent it on a screen. The only solution is to reduce the number of strokes, which you have to do carefully so that you don’t inadvertently alter the meaning of the character.

In the past, stroke reduction involved embedding bitmaps for each Japanese character, an incredibly time-­consuming process given the sheer number of characters. One company reportedly spent 50 person-years to create a new Japanese computer typeface.

By contrast, the new font, called Meiryo and designed by Eiichi Kono, Verdana creator Matthew Carter, and Japanese font company C&G, took just two person-years to develop. The font team was able to work so quickly because they applied the basic concept of automatic hinting to the task of stroke reduction. They still tuned the 3000 most frequently used kanji characters by hand, but for the next 6000 or so characters, they used software tools to do the initial hinting and stroke reduction, followed by manual adjustments. The 12 000 or so least-used characters were completely hinted by computer [see figure, “Different Strokes”].

Of course, improving screen resolution two- or threefold would make a lot of these typographic enhancements less necessary. But for the reasons cited before—the power needed even to double the pixel density, the cost of making denser screens—that’s not likely to happen soon. Short of increasing the raw number of pixels per inch, what can you do to add clarity?

Early computer fonts assumed that pixels were either on or off, and the result was that their letters, formed from lots of tiny black squares, had a jagged look. To fill out the lines, font developers started adding slightly lighter squares at the edges of curves and diagonals, a technique known somewhat cryptically as antialiasing. Viewed close up, the lines actually appear a little blurry, but at a normal reading distance, the shaded pixels trick the eye into seeing what it thinks should be there: smooth continuous lines.

When color LCDs began to replace CRTs, Microsoft developers realized they could take antialiasing one step further. If you hold a magnifying glass up to a color LCD monitor, you’ll see the rectangular red, green, and blue subpixels that make up each pixel; a 5-by-5-pixel grid contains 25 pixels but 75 ­subpixels [see figure, “Color Coding”]. When turned up to maximum intensity, these colors trick the eye into seeing a white background.

Just as antialiasing involves manipulating the intensity of individual pixels, type developers figured out a way to manipulate the intensities of individual subpixels. To render a line that is only a fraction of a pixel wide, they illuminate only the appropriate subpixels—in effect, increasing the text resolution. Microsoft introduced this technique of subpixel rendering in 1998 under the name ClearType.

The latest version of ClearType, included with Windows Vista, pays attention not just to individual letters but to the spacing between letters. Previously, with “reading” size text of 10 or 12 points, we could place either 1 pixel in between letters, which was often too little, or 2 pixels, which was often too much. Using the extra resolution in the subpixels, we can now have fractional spacing, which improves the evenness and symmetry of the entire page [see figure, “Trading Spaces”].

There are other approaches to onscreen type, of course. While Microsoft stresses hinting to improve onscreen rendering, Apple and Adobe have focused on making the onscreen text as faithful to the printed output as possible. Instead of hinting letters, which slightly distorts the letter shape, they perform antialiasing on the letter outline, with slight stem-weight adjustments. The result is that when you look at a page of text onscreen, the text will have a very smooth, even appearance, much like the printed page. The trade­off is that the individual letters are less crisp and therefore more difficult to read onscreen.

Ultimately, to make reading onscreen truly equivalent to reading from the page, you need to solve the problem of portability. No one wants to be tied to a desk or have to lug around a laptop just to do some light reading. People want the freedom and flexibility to read lying down on their sofas, standing up in the subway, or while smearing cream cheese on a bagel in their breakfast nooks. Developments in tablet PCs, electronic books, and electronic paper show promise, but weight, screen resolution, and power consumption still have a long way to go.

Sony’s Portable Reader, for instance, is a lightweight electronic book device that relies on an e-paper display from Cambridge, Mass.–based E Ink Corp. Unlike an LCD, it can be easily read even in bright sunlight. Because the display draws power only when the image changes, power consumption is low.

But e-paper can’t display moving images or colors, so it’s mainly suited only for niche products like the Reader and similarly static applications. Laptops, cellphones, and other products will likely continue to use LCDs for the foreseeable future.