Comment by maxk42
3 years ago
Aside from licensing, usability studies show that serif fonts are more legible than sans serif fonts, so I have no idea where they came up with the idea that somehow they're harder to read for people with disabilities when all prior research I'm aware of has shown the opposite.
Yes! This a thousand times. Many universities I know now discourage sans-serif fonts for screen reasons and prefer Arial because one study showed a slight preference for those who were dyslexic. The evidence is far from strong [1] and frankly I feel like doing a meta-analysis of it in my spare time. I really, really, really dislike Arial and find it harder to read than others.
[1] https://scholar.google.com/scholar?q=dyslexia%20arial%20font...
I'm puzzled why sans-serif fonts would be discouraged for screen reasons. My understanding is that serif fonts are more visibly degraded than sans-serif fonts on low-resolution displays.
Displays are no longer low-resolution.
2 replies →
I don't believe that is true anymore, I've seen studies that show that there is little difference between the two.
It should be noted that older studies are all about paper and signs, while newer ones tend to focus on screens. There are some important differences between the two esp. when it comes to low-DPI screens (which are sadly still typical for office desktops).
I vaguely remember something about there being a difference on printed paper vs screens when it comes to serifs. But also, aren't most aircraft controls labeled in something like Futura? I believe the Apollo program and other aerospace studies decided that was the most legible.
Lower-dpi screens, including the traditional ~96 dpi of classic monitors, makes serif-fonts less clear.
High-definition screens (such as the iMac retina display I'm looking at presently) and many e-ink devices hit 200+ dpi (200--300 is not uncommon for e-ink), which is the starting point for many laserprinters, though those can go to 600--1200 dpi. Monchrome vs. colour will show some distinction in clarity (monochrome effectively has about 3x greater resolution in displays given three-colour picture elements (pels).
My own experience is that I much prefer serif to sans-serif fonts for any substantial reading, though labels and titles may read better in a sans font.
State's decision here seems quite flawed.
The thought does occur:
Given that Calibri is Microsoft Office's default font ...
... and that Mobile has vastly overtaken Desktop in number of devices (by roughly an order of magnitude) ...
... then it may well be that both Microsoft and the State Department are optimising for reading on mobile. Which could be defensible, I suppose.
Though a far better practice would be to adopt a flexible document standard (say, ePub or HTML, though they're of course not unrelated), which will then adapt to appropriate media characterstics and, say, display in a sans font on small mobile devices and serif in print or large, high-def devices and displays.
(Note that "large" and "high definition" are not necessarily identical: a projection display is large but often not especially high definition. Likewise public signage and the like.)
Edit: As this 2012 NNGroup (Jakob Nielsen) suggests: in "Serif vs. Sans-Serif Fonts for HD Screens". Though the piece notes that with high-definition displays of 200 dpi and above reasons to avoid serif-based fonts are mooted, there's still no clear argument for them either:
Almost all mainstream printed newspapers, magazines, and books use serif type, and thus people are more accustomed to reading long texts in this style. However, given the research data, the difference in reading speed between serif and sans serif is apparently quite small. Thus, there's no strong usability guideline in favor of using one or the other, so you can make the choice based on other considerations — such as branding or the mood communicated by a particular typographical style.
<https://www.nngroup.com/articles/serif-vs-sans-serif-fonts-h...>
(I note that as a strong fan of serif-based fonts for long-form works myself.)
Reading a control label is very different from reading a paper or onscreen memo.
Apparently the serifs don't work as well on OCR and screen readers.
https://twitter.com/John_Hudson/status/1615486871571935232
Why would you build a screen reader without targeting the most common font of the last thirty years? That's like making a license plate reader that has trouble with white license plates.
I thought the same thing. Surely it must be a function of the prevalence of these fonts though. In a hypothetical world where 90% of text you read is in sans serif I'd have to imagine that this would tilt the readability study results in sans serif's favor. I wonder if the studies attempt to control for this somehow?