When we read, we don’t have to think about recognizing the shapes of letters or words. We don’t have to think about jumping to the next line when we reach the end of the one we’re currently on. The process needs to take place automatically, below our conscious cognitive level, and good text is set up to allow that to happen.
Paragraph 2. 550 years of text mutation has taught us how to compose text so the human visual system can read it as easily and comfortably as possible. One of the techniques that really helps is to justify columns of text. It’s much easier for the eye and brain to tell when they’ve reached the end of a line if that always occurs in the same place. But there are problems with doing this on the Web.
Repeating the same text was deliberate. The first example’s set “aligned left”, the second’s justified. And here’s the problem with doing it on the Web. It’s very subtle, so you’ll have to bear with me.
(My writing style makes it hard to show this dramatically. As a former journalist, I tend to use plain language and short sentences, so there are not too many long words or technical terms).
Justification should make text easier to read, according to the theory. So how come Paragraph 2 – the justified one – somehow feels a lot less comfortable?
There are some subtle things going on here. Justifying text means you have to evenly distribute the spaces between the words. And that means word-spacing varies from line to line. In my writing style, it doesn’t vary by very much. But even that little is enough to interfere with the scanning rhythm.
How this is fixed in print is by hyphenating words to keep word-spacing more constant. It doesn’t need to be exactly the same on each line, as long as it doesn’t vary by too much.
At the same time, with all the lines of equal length, typographers have found it’s a good idea to insert a small indent in the first line of each paragraph. And that means you don’t need space between paragraphs – which not only saves space but eliminates all the “two-line reverse saccades” needed to get to the start of each paragraph.
The Web supports justification. But no-one’s doing hyphenation on justified text- so no-one uses it. They may not understand why they don’t like it, because the differences are subtle, as I said. But they don’t.
It’s like art. We may not know about it, but we know what we like. Even if we know nothing about type or text, our automatic scanning system knows the difference between good and bad.
A few years ago, my friends in the ClearType group and I showed Bill Gates two example pages: one well-set, the other other badly-set.
“It’s obvious this one’s better,” Bill said, “but why is that? We need to put a lot more science into this”.
There wasn’t much good reading research going on at that time – especially into reading on the screen. We hunted around for people who could execute research projects for us, and realized there were no real “centers of excellence”.
How these normally happen is that a professor or senior researcher gets interested in a particular topic. They get funding for research, and start some projects. Their graduate students help as part of their studies. Then they get interested, and develop some expertise. Often, they end up as full-time researchers working in the same department in which they studied. And then you get a body of experts in one place – a center of excellence.
We set about trying to help grow some new centers of excellence in reading research by providing funding. We hired a cognitive psychologist, Dr. Kevin Larson, to run the program. Kevin still works in the ClearType and Readability Research group at Microsoft (they’re part of the Windows team). He’s done fantastic work, and we now have plenty of universities who can do this work, in areas like ophthalmology, eye-tracking, modeling the human visual system, and so on, and we have an ongoing research program.
The illustrations in my blog were provided by Kevin.
Kevin’s helped us put a lot more science behind the Mystery of Reading,