Wednesday, January 06, 2010

Typesetting myths you should have gotten over by now

Having worked with and studied type for half a century, I have seen fads come and go. I have seen a lot of badly designed experiments by psychology undergraduates passed off as “research” into type readability. And I have seen a lot of opinion on the part of graphic design instructors passed off as fact.

But in the Dark Ages of my youth, type was set by compositors. Since the advent of desktop publishing thirty-odd years ago, people with no training in typography have been putting themselves out as experts. And that has caused the sewers to overflow, releasing all sorts of stuff into the mainstream. With the more recent advent of digital book printing, all sorts of people are typesetting their own books—or trying to. And when they dip their nets into that mainstream looking for tips on how to do the job right, they often come up with—well, you get the idea.

Naturally, I’d much rather see self-publishing authors avail themselves of my services or those of another professional compositor than do it themselves. But if you are a diehard DIYer, here are a few suggestions from a grumpy old typesetter on Internet myths you should ignore:
  • Don’t hyphenate; it slows reading and reduces comprehension.

    FALSE. If you are publishing an edition specifically for groups with impaired reading ability, you should avoid hyphenation. For normal readers (the large majority of the reading public for trade fiction and nonfiction alike), a moderate amount of hyphenation improves readability when compared with the alternative of having letters and words badly and unevenly spaced. You do want to avoid having more than two consecutive hyphens or, in narrow columns, three consecutive hyphens, but there is no reason to turn hyphenation off altogether.

  • Don’t justify text; it forces uneven word spacing.

    FALSE. Using modern page layout software such as InDesign, you can keep word spacing within an acceptable range, minimize (but not necessarily eliminate) hyphenation, and retain conventional justified text. Readability research confirms that normal readers continue reading longer and retain more of what they read when text is justified normally than when it is set ragged right.

    But note: Hyphenation is not available in HTML, so justified text looks terrible on the Web. For online use, such as here, stick with ragged right.

  • Tight type looks better than loose type.

    FALSE. Tight type increases reader fatigue and slows reading. If you look at books set on Linotype (hot metal) machines from half a century ago, you’ll see that the standard setting (governed by the mechanical limitations of the technology) was considerably more open than the default kerning values for modern digital fonts. Stay with the default or, if you want to be kind to your readers (especially at smaller point sizes), open the type up by a fraction. This is a good trick for situations where an old-fashioned appearance is desirable too.

  • You can typeset a perfectly fine book using Microsoft Word.

    FALSE. You can reproduce the Sistine Chapel using a paint-by-numbers kit, but it’s going to take a long time and you’re not going to fool anyone. Word is a word processing program, not a typesetting program.

  • All you have to do is import your Word document into InDesign and you’re done.

    FALSE. Composition remains an art that requires human intervention to produce good results. Learn to look at a page critically and to apply adjustments that improve the reader’s experience. Watch for ladders, rivers, pigeonholes, tight and loose lines, bad breaks, widows, orphans, and unbalanced spreads. Eliminating all of those at the same time is what separates the pros from the wannabes.


L. Diane Wolfe said...

I agree wholeheartedly on all of those points!

Good post, Dick.

Katharine Wiencke said...

What are pigeonholes?

Dick Margulis said...

Hi Katharine,

Pigeonholes are overlarge word spaces that don't align vertically like rivers but make the paragraph look a bit like Swiss cheese. About the only time you see one in books set with modern composition software is with a long URL that doesn't break in a convenient place and leaves a gap in the line where it begins. In narrow columns, such as in newspapers or in the back of magazines, you can get them in justified type with any long word that's hard for the software to hyphenate automatically, such as an unusual proper name or a foreign word. The problem can generally be solved easily enough if the compositor is paying attention.

JFBookman said...

Thanks, Dick, great post. It's amazing how persistent some of these myths are. Remember the "tight type" decade? Painful. But articles like this are the best defense against typographic drivel. A great use for grumpiness!

Dick Margulis said...


Yep. Tight type was all the rage in the 1960s into the early 1970s. The Photo-Typositor freed advertising art directors from the tyranny of the metal type body, and they went nuts with it in headlines. Then the early photosetting machines let them do the same with body text.

I walked into an office a few years ago and noticed the brass letters of the company name mounted on the wall behind the receptionist. Optima bold kerned tight. I said, "1969?" The receptionist looked up and said, "Pardon me?" I said, "The company has been here since 1969, right?" She had to check, but she confirmed I was right.

Unknown said...

I wonder if you could point me in the direction of some reliable research about these issues? I have a professor who's got his own style sheet. He's very strict about it. But I suspect that many of his rules (sans serif fonts for body copy, two spaces after every period, etc.) are all mixed up when it comes to readability. Thing is, he is not the sort of person who'd be impressed by unsourced claims in blog posts (neither am I, in fact).

So is there somewhere I can find a collection of links to various reliable (i.e. peer-reviewed) research on typesetting and readability?


Dick Margulis said...


Here's an overview of "research" on the subject of readability.

In the first half of the 20th c., type foundries funded a number of private studies on readability. If any of this material is still available anywhere, I'm not aware of it.

By far the greatest bulk of so-called research consists of very badly designed experiments conducted by psychology students (mostly undergrad, I think) for class credit. In reviewing a great many of these papers, I concluded that they didn't know what they were measuring or how to eliminate confounding variables. No typographer was ever consulted for any of these studies. They aren't worth the paper they were printed on.

There has been a good deal of expensive research done privately by the Starch Organization on behalf of their paying clients. None of this material is public.

To the best of my knowledge, there is nothing in peer-reviewed journals that is any better than the student work alluded to above.

That leaves exactly ONE resource, consisting of the results of well-designed experiments but unfortunately lacking in the requisite characteristic of being peer reviewed. I nonetheless present it for your professor's edification and enjoyment:

Colin Wheildon, "Type & Layout: Are You Communicating or Just Making Pretty Shapes"

Note that my focus is print, and that's what I'm writing about. Readability on a computer screen is a different kettle of fish and the characteristics of readable text are different from those of printed paper.