What Is Needed


Copyright © 2012 by John D. Berry
Originally posted at johndberry.com

Books are digital. This is not, strictly speaking, true; but it’s about to be, with a few honorable exceptions. Already today, pretty much all commercial books are produced digitally, although the end product is a physical one: ink printed on paper, then bound and marketed and sold. Already, the selling may be done as often online as in a bookstore. Already, the same books are being issued more and more in electronic form—even if, as yet, the e-books are mostly very shoddy in conception and execution.

But that will change. In order for it to change in a worthwhile way, we have to spell out just what form these books ought to take.

So what’s needed? How do we make good e-books? What should a good tool for designing and creating e-books look like and do? What should the result—the e-book itself—be capable of? And what should the experience of reading an e-book be like?

Last question first. If it’s immersive reading—a story or narrative of some kind—then you, as the reader, should be able to lose yourself in the book without thinking about what it looks like or how it’s presented. This has always been true for printed books, and it’s equally true for e-books.

But e-books present a challenge that printed books do not: the page isn’t fixed and final. At the very least, the reader will be able to make the font bigger or smaller at will, which forces text to reflow and the relative size of the screen “page” to change. That’s the minimum, and it’s a fair bet already today. But the reader many read the same book on several different devices: a phone, a laptop, a tablet, a specialized e-reader, or even the screen of a desktop computer.

For a real system of flexible layout in e-books and e-periodicals that might be viewed on any number of different screens at different times, what’s needed is a rules-based system of adaptive layout. I like to think of this as “page H&J”: the same kind of rules-based decision-making on how to arrange the elements on a page as normal H&J uses to determine line endings.

The requirements for this are easy to describe—maybe not so easy to implement. We need both design & production tools and the reading software & hardware that the result will be displayed on.

A constraints-based system of adaptive layout

The interesting problems always come when you have two requirements that can’t both be met at the same time. (For example: this picture is supposed to stay next to that column of text, but the screen is so small that there isn’t room for both. What to do?) That’s when you need a well-thought-out hierarchy of rules to tell the system which requirement takes precedence. It can get quite complicated. And the rules might be quite different for, say, a novel, a textbook on statistics, or an illustrated travel guide.

OpenType layout support. This means support for the OpenType features that are built into fonts. There are quite a few possible features, and you might not think of them as “layout”; they affect the layout, of course, in small ways (what John Hudson has called “character-level layout”), but they’re basically typographic. Common OpenType layout features include different styles of numerals (lining or oldstyle, tabular or proportional), kerning, tracking, ligatures, small-caps, contextual alternates, and the infinitely malleable “stylistic sets.” In complex scripts like Arabic, Thai, or Devanagari, there are OpenType features that are essential to composing the characters correctly. None of these features are things that a reader has to think about, or ought to, but the book designer should be able to program them into the book so that they’re used automatically.

Grid-based layout. It seems very obvious that the layout grid, which was developed as a tool for designing printed books, is the logical way to think about a computer screen. But it hasn’t been used as much as you’d imagine. Now that we’re designing for screens of varying sizes and shapes, using a grid as the basis of positioning elements on the screen makes it possible to position them appropriately on different screens. The grid units need to be small enough and flexible enough to use with small text type, where slight adjustments of position make a world of difference in readability.

Media query. This is the name used for the question that a program sends to the device: What kind of device are you? What is the resolution of your screen? How big is that screen? What kind of rendering system does it use for text? With that information, the program can decide how to lay out the page for that screen. (Of course, the device has to give back an accurate answer.)

Keep & break controls. These are rules for determining what elements have to stay together and what elements can be broken apart, as the page is laid out. This means being able to insist that, say, a subhead must stay with the following paragraph on the page (keep); if there isn’t room, then they’ll both get moved to the next page. It also means that you could specify that it’s OK to break that paragraph at the bottom of the page (break), as long as at least two lines stay with the subhead.

Element query. I’ve made up this term, but it’s equivalent to media query on a page level. The various elements that interact on a page—paragraphs, columns, images, headings, notes, captions, whatever—need a way of knowing what other elements are on the page, and what constraints govern them.

H&J. That stands for “hyphenation and justification,” which is what a typesetting program does to determine where to put the break at the end of a line, and whether and how to hyphenate any incomplete words. Without hyphenation, you can’t have justified margins (well, you can, but the text will be hard to read, because it will be full of gaping holes between words—or, even more distracting, extra spaces between letters). Even unjustified text needs hyphenation some of the time, though it’s more forgiving. When a reader increases the size of the font, it effectively makes the lines shorter; if the text is justified, those gaps will get bigger and more frequent. But there are rules for deciding where and how to break the line, and a proper H&J system (such as the one built into InDesign) is quite sophisticated. That’s exactly what we need built into e-book readers.

In digital typesetting systems, the rules of H&J determine which words should be included on a line, which words should be run down to the next line, and whether it’s OK to break a word at the end of the line – and if so, where. A system like InDesign’s paragraph composer can do this in the context of the whole paragraph, not just that one line. A human typesetter makes these decisions while composing the page, but when the font or size might be changed at any moment by the reader, these decisions need to be built into the software. In “page H&J,” where the size and orientation of the page itself might change, the whole process of page layout needs to be intelligent and flexible.

Up until now, in the digital work flow, the software’s composition engine has been used in the creation of the published document; the human reader is reading a static page. But now, with flexible layout and multiple reading devices, the composition engine needs to be built into the reading device, because that’s where the final page composition is going to take place.

It’s easy to create a document with static pages that are designed specifically for a particular output device—a Kindle 3, for instance, with its 6-inch E Ink screen, or a 10-inch iPad. I’ve done it myself in InDesign and turned the result into a targeted PDF. But if that’s your model, and you want to target more than one device, you’ll have to produce a new set of static pages for each different screen size and each different device. Wouldn’t it be better to have a flexible system for intelligently and elegantly adapting to the size, resolution, and rendering methods of any device at all?

Photo: a 17th-century Mexican handbook, about the size of a hand-held device, from the collection of the Biblioteca Palafoxiana, displayed during Typ09 in Mexico City. With ink show-through from the back of the page, which will probably not be a feature of e-books.

Emphasized Declarations


Declaration emphasized converted



Recessional: Different Tools for Different Jobs

stack of books

We’ve been talking about “books” without always being clear about what a book is. There are different kinds of books, intended for different purposes. We use a novel or a biography very differently from the way we use a dictionary, for instance. As a book designer, I’m used to thinking about this when designing a book: the optimum typography for a reference work may be quite different from the best typographic treatment of continuous text that we want to immerse ourselves in. A cookbook or a technical manual needs to impart discrete nuggets of information quickly and clearly, and to make it obvious what the relationship of those nuggets is to each other. An encyclopedia has to present an enormous trove of information in ways that you can dip into and locate the part that you’re looking for, and make the bit that you’re reading easy to absorb. A book that’s telling a story usually runs in a single direction, from start to finish; the typography ought to invite the reader into the page and then get out of the way, letting the author’s words go straight into the reader’s mind.

When we talk about the future of reading, and the future of books, we have to realize that this is a multifarious future: we’ll be reading one kind of book differently from the way we read another one, and what’s appropriate for one kind of reading isn’t appropriate for another.

As so often, in so many areas, there’s no one answer. We’ll need different tools for different jobs.

Is There a New Economy among Readers, Writers, and Publishers?

stack of books

We’ve been talking around ideas of a digital economy, starting with Jan Sassano’s suggestion that corporations like Twitter and Facebook make their money from the value that we, their users, give to them for free. How do we get rewarded for the value we’re bringing to them? (As a musician and ill-paid actor I once knew put it succinctly, “Money is such a nice way of showing your appreciation.”)

We’ve tried brainstorming ideas, though we keep getting caught up in digressions (which has been described as “the only way anything ever gets said”). There’s clearly no obvious answer to this question. Early this afternoon, after someone mentioned Socratic dialogue, I joked that maybe we should create the “Socrates app,” an app that would interrogate you and engage you in a Socratic dialogue. It would be a sort of ELIZA program with an edge. (No, I don’t have any idea how this would really work. But it’s an app! Apps all make lots of money, right? Right.)

Maybe I should crank up the Socrates app right now and have it ask us: “And how do you intend to make money?” Or maybe that’d just be too much like your skeptical parents shooting down your brilliant career plan.

We’ve tossed around ideas about control vs. influence (fame and celebrity of authors and other creators), thick data (deep-diving data about individuals), meaning vs. statistical quantities, and what Jan called the public experience of reading. That last led Wendy Ju to ask, “So, should publishers become special-events planners?”

We’re hardly the first to point out that writers today have to be self-promoters, and that live events have a major role in the ecosystem of publishing. The balance varies: a poet who is published by a small press may actually sell most of their books at live readings, while an economist or business consultant might give away their books as a sort of calling card when they do a public lecture or workshop, which is what they really get paid for. (Does Edward Tufte make his living from selling his books, or from presenting his one-day courses? Both, perhaps.) Academic authors often get their real reimbursement in academic credit and kudos, which translates into professional advancement at their institution, rather than through the pittance they’re actually paid (if any) for an academic book or article.

So is the future of publishing going to be live events? Public reading? Online interactive discussion, annotation, and response? Dan Gillmor said that he gets great value from the comments on his writing that come back to him from his readers, that in fact it adds to and improves his writing. There’s a big difference between applause and laughter at a live event and the approbation of comments on a blog or website, but they’re both response. And neither one pays for the groceries.

Is it possible to be a professional reader? If social media services ought to be paying us for participating in their game, should publishers be paying us to read their books? Or for allowing our responses to be measured, and aggregated and analyzed by the publisher? Maybe. How would that work?

Thinking About the Page

stack of books

The book page is an artifact of the delivery medium: it’s written or printed on a physical sheet of paper (or a carefully prepared sheepskin turned into parchment, or a section of a papyrus scroll). The two-page spread is simply an artifact of our codex form of book-binding, which always opens to two facing pages. But what’s a page in an e-book?

The onscreen page isn’t really the medium on which the content is displayed so much as it is a window into a part of the content. The part that’s being displayed is the “page,” but how much content fits onto that page depends on a lot of factors, such as the choice of font, the size of the fonts, and the space between lines. In the resizable, reflowable world of e-books, the “page” is constantly changing.

It also might change if you go from reading a book on your laptop to continuing to read the same book later on your phone; the screen size and shape can be entirely different. It’s not just a matter of scaling the whole page up or down in size. (Though if you’re reading a fixed layout, such as a PDF of a page from a printed book, then that’s exactly what you’re doing. Which is why it’s such an unsatisfactory way to read a book onscreen, unless the PDF has been designed for the format you’re using.)

Web designers and interactive designers sometimes refer to the “viewport,” which is the screen or the part of the screen that the content is displayed within. (When I say “the part of the screen,” I’m thinking of the way most web browsers and many apps impose a scrim of “chrome” around the live part of the page: buttons or frames or navigational tools, anything that’s not the basic content.) This is a useful way of thinking about it. Just as the printed text fits onto the physical page, the digital text fits within the bounds of the viewport.

Although a book of long, continuous text functions as a single flow from beginning to end, there are other kinds of books that involve a complex set of elements, both text and visual, that have to be arranged somehow with respect to one another. Think of a cookbook or a textbook. On the printed page, the book designer can arrange the elements appropriately, so the book is easy to use. The list of ingredients comes before the step-by-step recipe, for instance, and is clearly delineated as a separate element, or the footnotes and sidebars and numbered illustrations are set apart from the running text in a way that makes them easy to find when you want them, yet keeps them from getting in the way while you’re reading the main text. A good designer is always thinking about each visual element, including plain old paragraphs of running text, in relation to the space on the page and its edges.

How do we do that on a digital page that keeps changing? Can we design a complex “page” based on the size and shape of the viewport? Unless we’re going to restrict the reader’s ability to change things—font size, for an obvious starting point—then we’ve got to invent rules and patterns for how the elements on the page will behave when something changes. Creating the design and production tools that will give a designer that ability is one of the most important tasks ahead of the e-book industry.

In my experience of designing books and periodicals for many years, I’ve found that the most important single factor that makes a paragraph of text readable is the length of the line. Not the physical length (3 inches or 20 centimeters or whatever) but the number of letters and words that fit on each line. The usual rule of thumb in English is that a line of about ten words works well (with an “average” word of five letters plus a space), or 60–70 characters per line. Obviously that might change from one language to another, and it would be entirely different in a written language that uses ideograms rather than letters (Chinese, for instance), but I’m sure each language has its own rules of thumb. Also obviously, the ideal line length might vary from one writer to another. (I know that from experience. I once designed a book of essays by shamelessly swiping the page design of another book of essays by a favorite writer of mine, designed by an excellent book designer; but the pages I designed didn’t work as well. It took me a while to realize that the writer of my book habitually used words and phrases that were hard to break, so I kept getting too-wide spaces between words in the lines, despite my best efforts. And it wasn’t that he used long words; it’s much easier to find a hyphenation point in a word like dissimulation, for example, than in through or enough.) Rules of thumb are just descriptions of useful patterns; they always have exceptions.

I’m always amazed at people who look at websites by having their browsers “maximized” to the full width of a laptop or desktop computer screen. Unless the website has been designed to accommodate this, the lines of text are almost always much too long to be comfortable to read. The same problem comes up in printed business documents with inadequate margins; a page of 10-point Times New Roman spread across seven or eight inches is simply not easy to read. (It’s worth keeping in mind that Times New Roman began life as a newspaper typeface; it was designed to be read on newsprint in narrow columns.) Line length matters.

A smart digital page would adapt its layout to the size and shape of the viewport. [I wonder if we also want to give users more options for understanding their choices. People could voluntarily go through a short tutorial/quiz that informs readers of the contextual differences between, say, sans and serif, and how that can affect the reception of content. I also want every text to have an “author optimized” option, which is how the author/designer would like you to read it (including information about the best kind of device for the content). You don’t have to agree, but at least it could give the reader a less generic starting point.—AL] On a little phone screen, the text would be in a single column of short lines; on a laptop screen with a maximized window, the same text might appear in two or three columns, each of a comfortable line length. If the reader boosted the font size to a much larger size, the three-column page might rearrange itself into two columns. The smart page simply wouldn’t let the line length get too long for comfortable reading.

This is one of the reasons that we still need “pages.” A continuously scrolling page, like the archetypal web page, is only feasible if there’s only one column of text; otherwise, you’d have to keep scrolling back up to the top to continue reading in the next column. When people say they don’t like multi-column layouts on a web page, they usually mean this. Multi-column layout requires pagination.

“Why Am I Here?” the Man Asked

stack of books

What’s the purpose of a “sprint beyond the book”? I guess I’m here to find out.

I am both a book designer and a writer about typography and design, a typographer and an editor, an editorial designer and a reader. All those aspects of publishing and reading seem completely interrelated to me; I’m always surprised when someone concentrates on only one part or another. This “Sprint Beyond the Book” seems like an invitation to explore the ways in which all these parts fit together—not just today, or traditionally, but also in the future.

I’m not interested in just jettisoning the past of publishing and leaping bright-eyed into the unformed future without a thought. Nor am I interested in just re-creating the past of publishing in new forms, though I am committed to maintaining the strengths and the best traditions of book publishing and extending them in new ways.

E-books have followed the pattern set by other new technologies: like the earliest movies, which imitated stage plays, e-books have been imitating printed books, as though transferring a page spread from a printed book onto a screen would somehow be satisfactory. The reading medium is different. Besides obvious differences like backlit projected light rather than reflective light from a printed page (and the in-between forms of e-ink and its progeny), the most fundamental thing about how we read onscreen is that there are many screens: they’re not the same size, nor the same resolution, not necessarily the same aspect ratio (shape), and we look at them in a wide variety of physical situations (sitting at a desk, holding a tablet or laptop in, well, our lap, or peering at a phone screen in bright sunlight on a windy day).

As readers, we’ve gotten used to being able to change how what we’re reading is formatted—most obviously by increasing or decreasing the size of the letters. Just the act of changing the font size changes everything about the composition of the page; all those aspects of text design that a good book or editorial designer pays close attention to (and that a reader should never even notice, if they’re successful) get screwed up when the format is malleable.

What we need, then, is book pages that are smart, that adapt their design intelligently to changing circumstances; book designers of the future need tools that let us make decisions about how the design ought to change, so that we can set defaults for an e-book that will make it inviting and easy to read. (The user—the reader—may well change their settings from the defaults; but 99% of people never change the defaults at all, so they’d better be designed to work well right from the start.)

As the type designer Cyrus Highsmith has pointed out in his witty little book Inside Paragraphs (2012), the paragraph is the fundamental unit of how we read any extended text. And it’s the formatting of that paragraph that makes it readable. Starting with the invention of the word space in the Middle Ages, we’ve found ways to orchestrate and annotate the written word to make it easier to read. That’s what punctuation is all about; as Robert Bringhurst (1992) says, it’s essentially musical notation, telling us how to read the words. The typographer’s tools—line length, font size, color, spacing between letters and words and lines—all exist to help the author communicate and the reader hear.

So we’ve got our job cut out for us: to make up new forms for the book that preserve the values of reading and writing and publishing, while extending them into new areas and discovering what we can add to the art of cultural transmission.

Unbound Pages

Torn unreadable book pages

Originally published in The Magazine, Issue #13, March 28, 2013

Reading is a cultural act. What we preserve in writing and pass on through reading is our cultural knowledge, whether it’s instructions on how to change a lightbulb or a lyric poem written in response to someone’s death. For more than half a millennium we have relied on printed books for transmission of culture, along with an ever-expanding cloud of printed ephemera.

In recent decades, our dissemination of written knowledge has expanded without the need for physical printing. But we’re still learning how to read the unprinted word; and the people who lay out pages for readers are just now figuring out how to present those words in an easy-to-read form. That form isn’t always the same as the ones developed for books, magazines, and other members of the print family.

As the late Bill Hill liked to say, “No one ever asked us to upgrade to Reading Version 2.0.” Bill Hill was the co-inventor, with Bert Keely, of ClearType, software developed at Microsoft (one of my former employers) to increase the apparent resolution of type, making letters onscreen appear sharper. “We tend to take reading for granted,” said Hill, “since we learn how to do it at about five years of age, and we continue to use the same basic technique for our whole lives.”

Read Me a River

The mechanism of human reading hasn’t changed since we were puzzling out what shamans scratched onto tortoise shells or squinting at a bill of lading in cuneiform pressed into a clay tablet. Our eyes haven’t grown any bigger or shrunk any smaller, our arms still hold what we’re reading about the same distance from our eyes, and the size of the letters or other symbols that we’re comfortable reading for any length of time still falls within a narrow range.

If you’re reading this on an iPhone or iPad, you see the result of a whole series of decisions about how to present these pages in iOS in an attractive and readable form. It’s a static page—scrollable, but otherwise unchanging, except that the lines break differently if you turn your device from a vertical position to a horizontal one. (Different choices were made for the web, where there are more variables.)

The iOS app for The Magazine [where this article originally appeared] determines its articles’ typeface, set at a particular (but adjustable) size and with a particular amount of space between the lines. It also determines the margins around the text block, and all the other aspects of the appearance of the page.

But what is a page? In a printed book, that’s an easy question to answer: the page is one side of a sheet of paper. Or more precisely, the surface of one of the sheets of paper that, when they’re folded, trimmed, and bound, make up the book. The content of that page—text, titles, illustrations, captions, whatever—has to fit somehow onto that surface. The essence of a book is a lot of pages, bound together, with text sprawling across those pages in sequence, page after page.

The same can be true on a screen (any kind of screen, from a phone to a home cinema). Just as you’d lay out a page of text to fit on the printed page, you can lay out a screen “page” of text to fit on the screen. If your book is going to be read on several different kinds of devices, you can design the pages differently for each device (an iPad screen, for instance, versus a Nexus 7). But that’s still a static format: one page to one screen.

Another approach is to think of the screen as just a window onto a large page: you scroll up or down or right or left to see other parts of the page. We think of this as unique to computer screens, but in fact it reflects the way books were often composed before the format we’re used to—separate sheets bound down one side—became common. (That is called the codex format. Not to be confused with a software codec.) With an ancient scroll, the reader held the two rolls of the scroll, one in each hand, and read the page that was displayed in between. (Unlike what we see in mock-medieval movies when a proclamation is being read, book scrolls were held horizontally, not vertically.)

The “page” then was a block of text, written in relatively short lines and read, like a codex page, from top to bottom. The scroll was essentially a series of pages side by side, with the unseen pages rolled up on either side. The normal page of a scroll had lines noticeably shorter than a lot of our modern books, and the handwritten letters were usually larger than the printed letters we see today in, say, a newspaper or a mass-market paperback. But not by much.

On a screen, the fundamental design question is whether to make the page larger than will fit on a single screen, so that you have to scroll down or sideways to read, or to design a page that fits exactly into the visible area of the screen. (With a real physical scroll, the motion of “scrolling” moved from one page to the next, not down a single long page.) What you’re reading right now in The Magazine takes the former tack; you have to scroll down to read the rest of this article.