–Geoff-Hart.com: Editing, Writing, and Translation

—Home —Services —Books —Articles —Resources —Fiction —Contact me —Français

You are here: Articles --> 2009 --> Typography 101C: the role of typeface choice in making text readable

Vous êtes ici : Essais --> 2009 --> Typography 101C: the role of typeface choice in making text readable

Typography 101C: the role of typeface choice in making text readable

by Geoff Hart

Previously published as: Hart, G. 2009. Typography 101C: the role of typeface choice in making text readable. Intercom February 2009: 48–49.

In two previous articles (Typography 101A and Typography 101B), I discussed theoretical typographic considerations, but in practice, most of us encounter these issues at a far less abstruse level: that of typeface choice. Before proceeding, let's define our terms. A typeface represents a series of guiding principles for how the shapes of individual characters arise. Although the same character in any two typefaces is more similar than different (which is what lets us recognize it in both typefaces), the typefaces themselves are as distinctive as the faces of two friends. In contrast, a font is the specific implementation of a given typeface. Like fraternal twins, any two Garamond fonts are recognizably the same typeface, yet prove to be clearly different upon close inspection.

Serif and san serif type

We use two major typeface classes in technical communication: serif, a word that refers to the small flourishes (serifs) at the ends of the strokes that form each letter, and sans serif, a French phrase meaning "lacking serifs" (Figure 1). Serifs increase the distinctions between two characters in a typeface, thereby increasing legibility; sans serif characters lack these distinctive touches, theoretically making them less legible. In practice, modern sans serif designs use tricks such as variations in stroke width or exaggerated differences in character shapes that achieve the same result as serifs. A third category, ornamental type, serves primarily esthetic goals, though better designs also remain highly legible. A readable typeface is first and foremost legible, which is why "boring" old designs like Times remain popular: Times is space-efficient and highly legible across the range of sizes for which it was designed.

Figure 1. Serif and sans serif typefaces.

Prevailing wisdom (a phrase that always suggests the need for caution) states that serif fonts are best for body text because they're easier to read. Though this may be true under laboratory conditions, the increased legibility has little practical significance compared with a sans serif font designed for legibility. Many European and Nordic audiences are so familiar with sans serif type they can read it as fast as North Americans read serif type. Well-designed studies have unequivocally demonstrated statistically significant differences in legibility between typefaces, but I'm aware of none that demonstrated functionally significant differences without an obvious difference in inherent legibility. This is because many typographic studies are seriously flawed:

They compare typefaces that differ obviously in legibility, thus predetermining the results. For example, some compare an optimally legible, skillfully typeset serif typeface with a suboptimal sans serif typeface (e.g., one designed for headings, not body text).
They control test conditions so tightly to exclude extraneous factors that the findings apply only to the specific test conditions.
Conversely, they fail to control factors such as line spacing and perceived type size, meaning that these factors are being compared, not the two typefaces. For example, many studies compare two typefaces at the same numerical size, even though the typefaces then differ greatly in perceived size (Figure 2).

Figure 2. Typefaces can appear obviously different at the same nominal point size. Note that Garamond appears shorter, and Arial taller, than Times at the same point size.

Given access to a large type library, it's trivially easy to choose legible and illegible examples of both serif and sans serif typefaces. It's equally easy to set type so badly that even the most legible typeface becomes unreadable. Clearly, the presence or absence of serifs alone does not determine legibility. Many other typographic factors are involved, including those discussed in my two previous articles. In most cases, familiarity with and practice reading a typeface prove more important than theoretical considerations, and it's more important to choose a typeface to support a specific goal and to typeset it properly (see Audience and usage considerations).

Differences between media

Sans serif type is widely believed to be more readable than serif type for online text. This was unquestionably true in the early days of computing, when the display resolution was too low to clearly display the fine details of serif typefaces. Slab serif typefaces were a common solution because their thick serifs, which are often oriented vertically or horizontally, were more likely to be as wide as the pixels used to display them, and more likely to align with the horizontal and vertical lines of pixels on the screen. Thus, they were more likely to remain legible at low resolution.

With modern displays, resolution is high enough to legibly display even type with finely detailed serifs and varying stroke widths. (This is less true for devices such as cell phones and car navigational displays, with lower-resolution screens. For these displays, we must carefully test typefaces to ensure they remain legible.) Modern operating systems also offer font-smoothing technologies such as "antialiasing" that improve the legibility of all type: ClearType for Windows and Quartz for Macintosh. Interestingly, though many love these technologies, others (including me) find that they actually degrade legibility, and disable them as fast as possible.

This is an important reminder that whatever the theoretical advantages of a particular typeface, personal preferences can prove to be far more important. For example, I've been doing onscreen editing so long using some variant of Times that working with sans serif fonts dramatically reduces my productivity. Though it's tempting to prescribe typefaces based on rules of thumb (e.g., use sans serif online) or our own prejudices (more rarely, based on usability testing), it's wiser to let readers choose their own typeface and size where the technology permits this flexibility. Cascading style sheets (CSS) technology allows us to tightly control Web typography, while still letting readers modify how their browsers interpret our definitions. For example, most Web browsers let us increase or decrease type size until it is large enough for comfortable reading, and let us choose the default fonts and sizes the browser will use to display text when the Web site doesn't specify these characteristics. Some (e.g., Apple's Safari browser) even let us replace a poorly designed style sheet with our own style sheet.

Audience and usage considerations

The audience's nature must also influence our choices, since different contexts require different solutions. For example, advertising copy or material intended for young, trendy audiences require self-consciously "cool" typography, whereas the need to pack maximum information into minimum space without sacrificing legibility explains the tiny, dense text in phone books. Novel designers emphasize readability so readers can zip through stories at maximum speed, whereas wedding invitations are short enough that reading speed is irrelevant and visual elegance is far more important. Older readers tend to prefer larger type than younger readers will accept, although it's important to remember that visual acuity, not age, is the important factor that determines this preference.

Several studies have demonstrated convincingly that, at least under laboratory conditions, readers attribute certain characteristics (e.g., formal vs. casual) or even personalities (e.g., "fun") to typefaces, and it's tempting to make design choices based on these studies. Unfortunately, even though these studies offer intriguing insights into how people interact with type, perceptions of type remain intensely personal and subjective, and it's dangerous to overgeneralize study results. I've watched with amusement the recurring flame wars that spring up in online discussions between those who consider Times an elegant typeface and those who abhor it, as if typeface choice somehow had a single right answer.

Those of us who love typography sometimes forget that most readers never notice our typeface choice until it interferes with their ability to read. I've seen designers make foolish type choices purely to appear edgy or cool; for example, I abandoned Wired magazine and didn't return for many years after grappling with their early, occasionally bizarre, design choices. Several years ago, the editor of one of Montreal's "edgy" weekly tabloids commented that during their latest redesign, they’d spent months picking fonts and fiddling with the typography—and that in the end, nobody noticed. Or possibly nobody complained; I gave up reading their paper when the type size dwindled into illegibility.

Typeface choice is a balancing act

By now, it should be clear that choosing a typeface isn't something we can do by following a single set of rigid proscriptions, such as avoiding sans serif body text in print. For many purposes, default choices such as the typefaces that come pre-installed on all modern computers (e.g., Times New Roman and Verdana) are "good enough". Paying attention to the typographic details I've discussed in previous articles is more important than focusing on single aspects such as the presence or absence of serifs. Moreover, how well a text satisfies the audience's needs depends far less on typography than it does on our writing skill, the complexity of the subject, and the quality of the display medium.

Nonetheless, well-designed typefaces that are typeset based on an understanding of how audiences read and the limitations of the display medium remain important. If we pay attention to these factors, “good enough" usually is good enough. The best information designs are invisible to readers, and as soon as the design attracts attention away from what we're trying to say, we risk failing to communicate. Probably the best way to confirm that our typographic choices were effective is to ask our audience to use what we've designed: if nobody complains about the typography, our design probably succeeded.