In the mid-1990s, the education world first saw digital content in classrooms in the form of CDs in the backs of books. Publishers threw in these digital “additives” to make their print products more competitively attractive as personal computers began to show up in classrooms. While the extra content was in a few cases genuinely dynamic and instructional (like videos or games), more often they were simply paper ancillaries printed to the screen instead of paper: PostScript or WordPerfect files that were to be printed out by the instructor as needed. Worksheets that had for years been printed, then shipped, were now shipped, then printed.

On the surface, digital textbooks held out an intriguing two-fold promise: better content for educators and cheaper delivery for publishers. Educators saw visions of interactivity, collaboration, and dynamic media. “More exciting content” promised greater interest and engagement for students. For publishers, though, direct digital delivery meant something further as well; it meant the potential for no printing, warehousing, and shipping of books. A digital story might get faculty adopters to look at a product they would not otherwise consider, and that was attractive, but of at least equal interest was the potential for publishers to get out of the ever-more-expensive pulp, print, warehousing, and diesel fuel businesses they had always considered more necessary evils than core strengths.

By the late 90s and early 00s, educators and publishers understood that the personal computer was on an inexorable march toward the classroom, and that more sophisticated digital products were going to be needed. Schools were getting “wired,” computers were getting better and cheaper, and faculty began to acquiesce to computer training.

During this period, big computer companies in particular began to throw their marketing weight behind large-scale student computing programs. These programs ranged from laptop “carts” in classrooms—where students checked out laptops from time to time—to full “one-to-one” computing initiatives, where each student received her or his own laptop for use around the clock. In 2001, for example, Henrico County, Virginia, and Apple Computer introduced a massive one-to-one program for all middle school students across the county. Every student in every middle school was provided a personal laptop. Eventually, high schoolers were added. During the same period, IBM (later Lenovo) launched similar initiatives in higher education, providing large-scale programs at Wake Forest in 1996 and at UNC–Chapel Hill in 2001. Coupled with steadily advancing network infrastructure, student computing initiatives appeared to be the requisite technical groundwork for e-textbooks.²

In practice, though, the content was not forthcoming, and what the market learned was the first rule of digital content initiatives: implementation ruins everything.

For schools, different products needed different computing platforms and different add-ons. Some software ran only on Windows platforms, for example, and a given classroom might have only an Apple computer; some programs might need a CD-ROM drive or an Adobe plug-in that the school computers did not have. In most places, it was still too expensive to provide a majority of students with access to a computer for any extended period. Computers labs—the typical approach at this time—were a good place to teach basic keyboarding skills, but they were not a viable distribution point for mission-critical course materials like textbooks. In reality, most students during this period got no more than a few minutes per week at an actual computer.

At the same time, content providers found they faced myriad problems taking advantage of the digital channel. First, most publishers had never before signed author contracts that allowed for digital distribution, so they were unable to provide the content even if they wanted to. Renegotiating these contracts was a difficult and expensive prospect. At the same time, there was no revenue model that provided return on investment (ROI) for the digitization and licensing efforts. If a publisher developed an elaborate digital version of a text, the cost was substantial, and there was no new incremental revenue to pay for it. On the contrary, in most cases the market expectation was that unit price would decrease as products migrated to digital. Publishers were therefore going to make less per unit, were in theory only replacing print sales with digital sales, and were being asked to nonetheless invest more per title for content development.

Worst of all for e-textbooks, it was during this same period that Napster came along and demonstrated just how disruptive digital change could be for intellectual property owners. Publishers became highly selective about releasing content into digital form, and when they did release it, they often insisted in access restrictions (like no copying or printing) that rendered the products all but useless to active, collaborative learners.

While a few interesting digital content products emerged during this period, the complexities of digital content development, the establishment of a meaningful marriage between the digital tools and instructional aims, and the complete lack of uniformity in the technical environments of schools made implementation much more difficult than many had hoped. Implementation ruined everything.

Nonetheless, computer access and technical infrastructure continued to improve in most schools, and by the early-to-mid-00s, a few independent platform providers had emerged to help publishers scale their digitization efforts and deal with ROI and digital rights management (DRM) issues.

VitalSource Technologies, for example, introduced a full curriculum product to several US dental schools in 2000.³ Every student in these programs received a laptop preloaded with every title they would need for their four years of study—along with key volumes of professor-created documents—and all titles were regularly updated through graduation. Platforms like VitalSource applied search and annotation tools across all titles, but also provided sophisticated DRM, thereby protecting content and helping publishers scale their digitization efforts. The program has expanded in recent years into medical education, and by 2006 one in three dental students in North America obtained their entire textbook course of study through this program.⁴

In another example, Elsevier Health Science announced an innovative “paper-plus-electronic” product in 2005 called Evolve Select, now called Pageburst.⁵ This product, originally for nursing students, but later expanded to other health science professions, provided students with both print and digital copies of texts. Students were able to read print if they liked, but could also use software search and sharing tools as needed. O'Reilly and Pearson teamed up in 2001 to provide the SafariX platform, which delivered online textbooks for technology disciplines, and that program also continued to expand during this time.⁶

Most of these niche successes, though, turned out to be segment-specific and small in relation to the overall market. This led the industry to face the second rule of digital content initiatives: people can't buy what they can't find. Research carried out by Blackwells UK Ltd. in 2010 suggested that in mid-decade the primary hurdle to digital adoption in the market was simply the lack of available core reading list titles in electronic form.⁷ This did not change until late in the decade when the five largest publishers in higher education banded together to create a common distribution platform: CourseSmart.

Initially focused on instructor sampling, CourseSmart used the aforementioned SafariX technology to deliver online e-textbooks that were exact digital copies of the print versions. Soon after launch, it added a download option based on a fully branded version of the VitalSource platform and began marketing directly to students as well. CourseSmart made available to instructors and students almost the full array of frontlist titles in almost every subject area, and so for the first time students who were interested in trying digital versions had a reasonable chance of at least locating their desired titles.

Once the titles were available in one format, however, it was much easier for publishers to make them available in many formats and digital channels, so e-textbook options began to show up in a variety of channels, like MBS Directs’ textbooks.com, Follett's CafeScribe, and Ingram's VitalSource.

From a business standpoint, then, the contest shifted from title availability to format quality, usability, and business models. CourseSmart titles were available for rental only, and for either online or offline use, not both. The rental price, however, was uniformly 50 percent of print list price. Other providers offered joint online-offline access and rental-or-perpetual purchase options, and price point fluctuated with access terms. Some platforms also offered reflowable products that performed better on emerging handheld platforms, and some began to embed multimedia objects and assessment engine links into their books.

Also, the publishers themselves continued to try to establish their own individual digital offerings. Cengage purchased the SealedMedia technology to deliver its iChapters product (recently the name was changed to Cengage Brain), which allowed students to purchase only the individual chapters they needed. The Pearson MyLab suite of offerings complemented textbooks with online tools like active assessments and multimedia animations. At the same time, John Wiley and Sons developed WileyPLUS and McGraw-Hill developed its Connect system to similarly complement its textbook offerings. Elsevier's Evolve and Student Consult products complemented its print and digital textbook options in the health sciences. Each of these products is delivered differently with unique pricing options.

While publishers continued to wrestle with rights management issues, innovations like high-speed print scanner technology threatened to make digital printing and copying a moot concern. Digital copies of textbooks could be made in minutes by anyone with a scanner and a print book, so heavy protection measures on digital files were becoming ever less relevant.

By 2010, the annual Educause ECAR study was reporting that nearly every college student in the United States owns both a computer and a phone, and that 80 percent own a laptop. Not only was the relevant inventory of product now available across the market, but the infrastructure was also in place—at least in higher education—for digital consumption to reach mass markets.⁸

At decade's end, there remained really only one question in many people's minds, but it was a big one: were students and faculty really interested in moving over to digital consumption of this content in the first place? One study widely cited in the press appears to put a fine point on the issue. The Book Industry Study Group's (BISG) Student Attitudes toward Content in Higher Education cited a study commissioned by the National Association of College Stores (NACS) and performed by OnCampus Research that reported 74 percent of US college students surveyed still “preferred printed textbooks.” According to BISG in its press release, students listed permanence, look, feel, and ability to resell printed editions as the key influencing factors. The survey is said to have further found that prohibitive pricing of new printed textbooks were driving students to seek more affordable solutions, either purchasing cheaper versions from online resellers or looking for textbook rental options.⁹ These findings appear compelling, but blogger Kent Anderson of The Scholarly Kitchen has pointed out that while the survey results were quoted in press releases, the results themselves have not yet actually been released. Anderson also pointed out that the conclusion quoted most often in the press—that students simply prefer print access to digital access—appears overstated. In a post from November 3, 2010, Anderson points out that economics and selection are referenced as the survey's key factors. For most students, a used print product is seen as inexpensive and readily available. Of course students prefer less expensive versions of a known commodity to an unknown product at an unsure price point, Anderson says. Confining the finding to being a strict “media choice” between print and digital does not appear warranted.¹⁰

Another key lesson that has emerged in recent implementations is the need for cross-publisher access platforms. While several of the largest publishers have invested heavily during this period in content delivery solutions that were particular to their own content, schools continue to see benefit from platforms that aggregate content. When unique, publisher-specific solutions are adopted on campuses, students must log into different systems with different credentials to use different applications with different attributes and different interface logic for different classes. This situation is analogous to each of the major car companies having its own proprietary type of gasoline that works only in its cars. In many scenarios the unique systems are acceptable, but in schoolwide adoptions, a cross-publisher platform had advantages. In a few cases, a campus or at least one department may be willing to standardize on a given publisher, and in those cases the problem goes away, but experience suggests that very few US campuses are willing to entertain such standardization, and so in reality, the market is better served in many cases by third-party e-textbook platforms that normalize features and functions and unify logins for content across all classes.

Finally, as the decade waned, along came Amazon's Kindle, and not too long after, Apple's iPad. Indications are that these devices are changing people's attitudes about reading on screen, and while the connection is not likely direct, the release of these devices in recent years has unquestionably coincided with growing sales of e-textbooks. Coincidentally or not, industry groups like NACS, MBS, and Simba have indicated rapid growth and rapid acceleration of growth in digital textbook sales during the period that these devices have entered the market.¹¹

Ironically, neither device has to date delivered more than a sprinkling of what most of us would recognize as “textbooks.” The Kindle's screen, being black and white, is not ideal for full-color textbooks, and the device does not offer the strong integrated web browsing, media support, or external application support (for things like Word and Excel), typically expected for study in most modern subject areas. The iPad, while certainly more developed in color, browsing, and application support, has more limited battery life and has yet to demonstrate broad title availability. In the fall of 2010, another big name in the book industry, Barnes and Noble, joined the fray by releasing an education-specific version of its Nook e-reader, called Nook Study. These device-centric reader solutions have elicited mixed reactions so far, but they will continue to improve and will certainly impact the market significantly going forward.¹²

In the meantime, a few companies are working with publishers to create a new breed of e-textbook products specifically for the iPad/iOS environment. Inkling and Modality, for example, have both identified iOS as a key content delivery platform in the coming decade, and both companies are working to make a broader set of titles available in their own apps.

By far the most common form of digital textbook today is the “page-fidelity” or “print-fidelity” product: exact screen renderings of the printed pages. These products rigidly maintain the layout of the paper version of the book, and indeed are most often built from PDF source files exported directly from the publisher's print workflow. More often than not, the PDF source is ingested into a third-party platform that applies some level of search and annotation functionality as well as DRM.

Proponents of page-fidelity e-textbooks point out they can be easily produced in great numbers through a single workflow. Production and quality assurance processes can be standardized and streamlined across multiple imprints, making it a cost-effective approach. These products do not therefore typically represent a substantial extra cost for publishers, and a very wide range of titles can be made available in the market very quickly.

In addition, the resultant products are familiar to students and faculty (they look like print books, have familiar tables of content and indexes, etc.) and are supported in most of the major DRM schemes in the market, such as those employed by Adobe's Digital Editions, CourseSmart, VitalSource, and CafeScribe. Best of all, they require little change in teaching behavior to be used in the classroom (so professor adoption is less stressful).

Detractors point out, though, that the products are fairly static, do not usually take advantage of the basic media and communication capabilities that make a computer so interesting in the first place, and are inherently inaccessible for the disabled.¹⁶ Worst of all, they require little change in teaching behavior to be used in the classroom (so there is no real classroom innovation, and learning is not necessarily enriched).

Other significant concerns include large file sizes (in download models files sizes can be hundreds of megabytes) and difficulty in integrating multimedia. Media objects may be linked to product, but with PDF-sourced documents, media objects cannot typically be embedded inline in the page.

In U.S. higher education, CourseSmart is a particularly informative case study. It implemented a page-fidelity product workflow in 2006–7 that scaled across not only imprints, but across multiple publishers. It was therefore able to grow its inventory of titles very quickly. These page-fidelity products have limitations as well, though. They do not typically introduce the new capabilities or enhancements that often make the switch to digital most attractive to many educators and learners.

“Reflowable” digital textbooks, on the other hand, maintain all the content from the print textbook, but often strip away or dynamically deal with components of page layout (see figure 5.2). These products—typically constructed from XML source files instead of PDF—have fluid line and page breaks. Here, lines of text wrap and/or resize as viewing windows are adjusted or as different devices cause readers to encounter different screen sizes. A reflowable e-textbook allows users to adjust font sizes and adjust windows to their liking without causing the entire page to resize. In many cases users can also designate their own preferred background colors for pages, figures, or box features.

Proponents of this class of product point out that reflowable texts provide a better experience on handhelds and other smaller device screens. Reflowable products also allow content creators to link or embed multimedia objects directly inline in the text, thereby presenting them within the context of the given chapter of study and making them more likely to be used. For smaller devices and for richer content, reflowable e-textbook environments would therefore appear advantageous (see figure 5.3).

Additionally, an XML-based reflowable product has to date proven to be a better option for disabled users. A document's XML markup allows for embedded tagging that describes figures, equations, and other nontext objects. The reader software uses this explanatory language to describe nontext items to visually impaired users. The more structured mark of XML also provides a clearer technical path for screen readers, Braille readers, and other external devices to follow for navigation, thereby improving windowing, menu advancement, and text-to-speech options (see figure 5.4). Schools and organizations dedicated to serving the disabled community and meeting their obligations under Section 508 of the US Federal Rehabilitation Act are likely well served by understanding the options provided by reflowable e-textbook product.

Section508.gov

www.section508.gov

The primary disadvantage of reflowable e-textbooks has been the investment cost needed to create product. The product is typically generated from XML source files, and these files are not typical products of the traditional print process. In fact, the XML standard itself has not been evenly applied across education publishers (to say the least), and there are wide variations from one publisher to the next both in how the specification is defined and how it is applied.

XML is a “markup language” and as such is simply a set of rules by which a document should be “tagged,” or marked, so that applications know how to display the content. Applications receive these tags as instructions (essentially, “make this word bold, that one italic,” “insert this picture here,” “start a new section there,” etc.), but various software platforms have different versions of the rules. The result is that without a common industry standard for markup, a content creator may have to tag a document separately for each potential channel. In theory, software translators can be created to automatically export to the various sets of rules, but in practice the process almost always involves some level of human decision making, and that in turn means an extra quality assurance (QA) step should be introduced for each channel. In recent years, publishers have paid external publisher-services companies, primarily in India, for this process.

Most of the large publishing houses have introduced XML into their internal workflow; however, it is typically more for archive and editorial control than for product mastering. By its nature XML is an extensible, or modifiable, markup language, meaning that each software provider can potentially ask publishers for a source file built to a different XML specification. Each different education channel has historically supported its own unique “flavor” of XML, so for publishers it has been expensive to export different flavors to the different reader platforms. The publisher would need to separately export (and separately QA) each version.

Unquestionably, XML-based product has many advantages in the market, so several years ago the publishing industry began to address the issue by creating an industry group to suggest standards for XML in publishing.¹⁷ This group, the International Digital Publishing Forum, or IDPF, has released several iterations of an XML and file-packaging specification called EPUB. While early versions of the standard were focused on the trade publication market, the most recent—EPUB 3—has been substantially upgraded to handle many of the structural, page layout, and metadata complexities inherent in academic publishing.

One promise of this new EPUB standard, with its more elaborate support for page layout, is a blending of reflowable and page fidelity formats into a single file workflow that will retain some measure of page fidelity in most circumstances, but also allow form-factor or device-specific adjustments when appropriate.

This new version of the standard does a much better job than previous versions of supporting style sheets, for example, and so should make it possible to create a richer layout. As an XML standard, though, EPUB-based products should maintain strong support for accessibility for the disabled. The EPUB 3 Working Group Charter specifically mentions expanding support accessibility, including use of the DAISY accessibility standards.¹⁸

Platforms like Adobe Digital Editions and the VitalSource Bookshelf support both page-fidelity and reflowable product. In the past, they have done so by allowing content creators to submit one type of source file or the other, and so the user would receive either the one type of product or the other. There has not been an option in the market for the user to “toggle” between types as they choose.

In the last year or so, the market has seen the emergence of e-textbook products with more interactivity and more embedded media. The traditionally leading platforms are working with publishers to embed video and links to external systems directly in their content, and new companies are taking advantage of emerging hardware platforms—like tablets based on the Android operating systems or Apple's iPad devices—to apply new levels of interactivity to books. Figure 5.5, for example, shows an inline video in an Elsevier Science nursing book in Elsevier's Pageburst application.

Inkling and Modality are two newer companies leading the way in creating books specifically for the iPad. Both companies substantially re-engineer the publishers’ source files to allow for greater user interactivity. While the new levels of functionality are exciting, the scalability of the approach is not yet proven. Many other, smaller companies are releasing titles as individual apps for iOS, but to date these applications have been primarily for trade and reference books and not textbooks.

At the same time, as the leading publishers in the space look for new ways to enhance their titles, they are also delivering new media-rich and interactive types of products. One such product released in 2010, Cengage's MindTap, is essentially a web-portal-based book that offers assignable and gradable quizzes, and what Cengage calls “adaptable learning paths” that adjust the ordering of units based on student performance and preferences.¹⁹ Going forward, the publishers will continue to balance the need to develop their own unique solutions with the need to participate in multipublisher implementations through third-party platforms.

In the market today, all the largest publishers have their own “internal” media-rich or interactive e-textbook products:

• Pearson MyLabs
• Cengage Brain (formerly iChapters)
• McGraw-Hill Create and Connect
• WileyPLUS and Wiley Desktop Editions
• Elsevier Health Pageburst
• Macmillan Dynamic Books

As publishers learned back in the 90s with title-specific CD products, though, title-specific media enhancement and interactive features can mean substantial investment on a per-title basis.