Books: Entering the Age of Glosses

Books: Entering the Age of Glosses

Summary: Here's the key to thinking about the future of writing, something straight out of the manuscript era: the humble gloss or "scholia," for those who prefer the Latin. They are the notes, in margins, footnotes at the bottom of a page (the standard starting around 1700) and later in the history of books endnotes at the back of the volume or in a separate appendix, that add interpretations, background information, commentary and definitions.


Here's the key to thinking about the future of writing, something straight out of the manuscript era: the humble gloss or "scholia," for those who prefer the Latin. They are the notes, in margins, footnotes at the bottom of a page (the standard starting around 1700) and later in the history of books endnotes at the back of the volume or in a separate appendix, that add interpretations, background information, commentary and definitions. Creating marginalia is an art made for the era of "crowdsourcing."

Steven Johnson's essay in The Wall Street Journal last week, about his e-book "'aha' moment" describes a number of new forms for the book. A worthwhile read. Johnson focuses on the unit of consumption, including the possible bundling of chapters of different books to facilitate sampling and the potential for a "global book club" about any word or sentence within a book, not just a focus on the whole book.

I think the deep change that will change what we think of as reading and "the book" will be the end of the idea of the finished book. It was only in the 18th century, when binders stopped adding extra pages for notes in the books they produced for individual customers and standardized binding, that writing in books became a taboo (despite many notable rebels against that trend), and before that using notes in books was a widely recognized form of social communication, according to author H.J. Jackson's Marginalia: Readers Writing in Books.

Last fall, I started researching the future of books and am now working on a book about the future of books, because a market research report about e-books would miss the real challenge faced by authors and publishers, along with their readers: How to see past the ideas we have about the artifact we call a "book." After the jump, I'll share an excerpt from the book that summarizes why cryptographic features, which could easily be tossed out with DRM, are the source of one of the most important potential reservoirs of value for publishers in The Age of Glosses, when readers share ideas through books and the idea of a finished document is not obsolete, but is increasingly rare.

Curation: Going beyond format

The assurance that you can reconnect to a backup of a digital document, such as a book, adds value that can be a significant differentiator in a world where content is supposedly free. Preventing viruses from coming along with a download is often offered as the benefit of paying for a book or song versus downloading a copy from an anonymous FTP or BitTorrent site, but that is small fish compared to the “curation” an online library gives book owners; the backed up library becomes a browsable collection, just as books on a shelf do, providing the satisfaction of ownership and the possibility of rediscovering a book without paying for it, again. On top of keeping copies for customers, Audible upgrades those books from older formats to new ones as they become available. Since Audible launched, it has introduced four upgrades, each providing better audio quality or compatibility with new listening technology—books are updated, so customers can get a better listening experience for free.

Audible’s new parent company,, has embraced this with a “digital collection” of the titles and products bought on the site. However, while titles may be downloaded multiple times, Amazon has not made electronic copies of books available across formats, so one must buy an e-book for Kindle and another that can be read as an Adobe PDF file on a computer. For example, in order to have access to a copy of The Cambridge Companion to Kant, a book I’ve read along with the works of the philosopher, I had to buy the paper edition ($29.50 retail), the Kindle edition ($28.42) and the “Amazon Upgrade” PDF version ($7.39) in order to access the book physically, through the eBook reader I own and the PC I use, respectively. Netted out, I spent $65.31, or 2.21 times the price of the paper book—and the copies are completely separate, isolated from one another. Notes made on the Kindle don’t appear in the PDF version, and neither of the electronic versions is linked to the paper book efficiently. The Kindle version reflows the book’s text without maintaining any page location, so I can’t make a note in the Kindle version and, when reviewing those notes, refer to a page in the paper book.

Format obsolescence, alas, seems to have been proved a viable business model by the movie industry, which has happily sold “versions” of the same movie on VHS, Laser Disc, DVD, and Blu-Ray, as well as for playback on computers and handsets, such as the iPhone. Hollywood sells a new copy at a different price, for each of the formats, rather than one ongoing connection to the film. Readers, thankfully, are not quite as gullible as movie-goers or, at least, they like to think they aren’t going to be suckered by publishers.

Now, think about that in relation to the way we purchase software. Many companies, such as Adobe and Symantec, to name two big firms, offer a choice of download or packaged copy (on DVD) of their products. Additionally, they offer ongoing access to download over a year or two for an additional fee. This lets the buyer decide to handle keeping a backup themselves or to, in essence, by insurance in the form of access to a downloadable copy of the application for an additional five or 10 dollars.

The software model justifies higher prices with “curation” of the customer’s software. A movie or book that is read once is, from the consumer’s perspective, worth less than something they want to keep and reuse. The DivX movie model, of selling one-time access to a movie, after which the DVD becomes unplayable, has not worked well. This is the case not because the distributor promotes one-time play, but because consumer ought to have the choice to extend their access without making a second purchase. If offered the chance to download and read a book once, or for a month, for a lower price than getting access to that book forever, which may carry a nominally higher fee, readers can choose their approach to a book—more importantly, if they change their mind and want to keep a short-term book, let them do it without downloading a fresh copy.

Here’s where cryptography can add a tremendous and novel value. Audible doesn’t maintain millions of different copies of the same book, each for a unique customer. A master file is encoded on demand and, by communicating with the Audible server, the customer’s copy can be updated with, specific to audiobooks, bookmarks and last-point-listened data. Granted, when Audible makes a higher fidelity version of the book available, the old copy on the customer’s player/PC must be replaced, but the cryptographic signature could also carry forward bookmarks and the most recent location in the book. However, note that in each case, the master copy is also common reference copy against which the bookmarks and last-time-listened could generate data that readers might use to find other who want to talk about the same bookmarked passages or who have read about the same part of the book.

The features of an e-book facilitated by cryptography, which is the same technology that underlies the much-despised “digital rights management” that many users protest against, are myriad. If only publishers would recognize that limiting reading with DRM was the greatest barrier to their enhancing the reading experience.

What’s so great about cryptography, without going into obscure technical detail? The simplest answer is that, using cryptography, we can identify an individual reader’s copy of the book, as well as their bookmarks, highlighted text, where they last stopped reading, and, if they book is linked to a network, what contributions the reader has made to different discussions about the book. Generally, publishers of music, movies and books have used cryptography to limit access to a single copy of a creative work, but the same technology opens new vistas if the limits are taken off and novel experience, even simple ones like mapping the location of a bookmark in an electronic copy of a book to the same location in a print edition, allowing readers to refer back and forth across different displays of the same title. Recall that Amazon’s Jeff Bezos said Kindle owners purchase approximately 1.7 times as many titles as they do printed books—at this early stage, it’s fairly safe to assume many of the books sold for Kindle use are copies of printed books the reader already owns, to extend the opportunity to read.

Cross-referencing of information in books, though it sounds like the dry purview of librarians and indexers, is the foundation of the conversation that is literature. Writers make literal references to other works in the main text, footnotes, endnotes, indices and bibliographies. Without a master copy of a work against which electronic copies that might be arranged in thousands of different layouts based on the size of the screen or application window in which it is displayed, e-books become unreferenceable. They also defy the primary benefit of networked computing, the ability to hyperlink, to make logical connections between texts that can be navigated. Because the e-book exists alongside the printed book, and will for the foreseeable future, such referenceability needs to extend from digital to analog versions so that all literary, scientific and professional annotations can be followed across formats. Cryptography can be of help here, too.

But where the cryptographic tricks become magical is in the potential for layering different versions of books or different “glosses” created by previous readers one atop the other, for use and comparison by readers. Imagine, for example, being able to select annotations to Joyce’s Ulysses made by a leading Joyce scholar or, in my ideal scenario, Joyce’s own notes or those of another author deeply influenced by Joyce: being able to juggle the different notes, bookmarks, highlights and hyperlinks provided by multiple commentators is the facility cryptographic tools bring to a text, because they identify each “layer” of the book, keep them in a manageable order, and, in conversations, even lets the reader engage securely, privately, even anonymously with other readers over the network from within the pages they are reading—regardless of how the pages are reflowed from one device to another.

This may sound like both nitpicking and wonkish thinking about how to read books, but consider a time when you were talking with a friend on the phone about a book or magazine article. The page number is essential to finding the common reference point. You said “turn to page 32 and look at the first line of the fourth paragraph” (which, if my memory serves, is the page of the paperbook edition of The Godfather, by Mario Puzo, that every kid in my junior high wanted to read—Sonny is having sex with the bridesmaid on that page). At the other end of the line, your friend said, “I found it.” That simply, you are literally on the same page. Were you reading on an e-book reader, the Amazon Kindle, for example, and your friend had a printed book, you have no common reference point, because the Kindle repaginates the book based on the size of the screen and the size of the font selected. Hence, if I turn to the beginning of chapter four of A World Without Time: The Forgotten Legacy of Gödel and Einstein, by Palle Yourgrau, my Kindle gives the “locations” 609-11, it does not present the page number in the printed book (Page 51 in the paperback edition from Basic Books). Even if you both have Kindles, but one of you reads in larger text, it is inconvenient to find a common reference point, except, perhaps, at the start of a chapter, if the book has chapter bookmarks. For example, the page location cited above appears when the Kindle font is set to “3” and locations 609-610 with font size “6”, which contains far fewer words than page composed in the smaller font. Granted, printed books change page layouts, as well, but they do it consistently, in separate editions that can be compared. E-books must do better.

In full flower, the e-book should be able to relate notes added in different editions to common locations, because master copies of editions can be compared and calibrated to allow consistent placement of annotations. If they did not, the result will be like getting a copy of a movie you know well that has the soundtrack out of synch with the image. E-books, because they support computational features, can harmonize information across generations. And not just so people can share the “good parts” of the The Godfather with their grandchildren or share notes with themselves at different ages as they revisit a book during their lives. These kinds of anchors make bibliographic annotation of any book possible, and will make entering the community evolving through a book a powerful addition to the reading experience our children expect when considering what they will read next.

Topics: Security, Hardware, Mobility

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • A Normative Approach

    Some very interesting insights here, Mitch. I'm excited about your work on the future of the writing, and I eagerly look forward to your forthcoming book.

    I do, however, have some trouble with the general framework from which you seem to be exploring the future of writing. Based on this entry, you are applying norms of print-based textuality to digital space, and this, I think, is a mistake. It's one that continues to be repeated in a variety of fields (not the least of which is my own--Rhetoric and Writing Studies--which continues to proffer hermeneutic norms of print-based textuality to all manner of meaning-making).

    This is to be avoided, I think, if we are ever to effectively explore the potential of writing in digital environments. Quite simply, print-based norms are not seamlessly analogous to digital writing, and remediating those norms brings along all of the baggage of book culture and applies it, ipso facto, to digital writing (even when we pretend or hope that it does not).

    In a certain sense, marginalia seems quite a suitable analogy to commenting, for example. But it's dangerous to extend that analogy into discussions of what writing <i>might be</i> in digital environments, because it so profoundly limits what we can see, reverse-engineering digital writing through the framework of book culture. Marginalia and glossing privileges <i>The Text</i>, not user interaction (see for example, C. Davidson, 1986 for more on marginalia). Glosses are furtinve, individual acts, isolated and contained. Blog comments, tweets, et al., are fundamentally <i>social</i> and public, privileging interaction between users rather than the hailed alphabetic text. It's not about <i>the
    book</i> alone, but about the writing that extends beyond.

    This is but one example, and the prevalence of normative, print-based approaches to writing in digital environments is overwhelming. So, more than anything, I stress caution in privileging such a framework.

    A way of seeing is also a way of not seeing.
    • A way of seeing....

      Thanks. I completely agree with your point. It is just an excerpt and
      from an early draft of this thing I am working on, and it is about books
      and the future of publishing, so I am sort of stuck with textuality as a
      starting point.

      Regarding textuality: For the purposes of communicating in text -- all
      of the modes of comment, tweets, etc, you refer to above -- I suspect
      we remain embedded in the textuality, or, at least, we're stuck
      referring to the textual quality of the social interaction. The
      interesting question is, when does non-textual annotation and linking
      change textuality into something new? For purposes of indexing,
      don't we always come back to the textual representation? I am not
      sure that, from our position vis-a-vis books today, we can imagine
      the "extends beyond" until we start by abandoning the standardized
      industrial artifact of "the book."

      To take a line from Jules in Pulp Fiction: But I am trying real hard....
      Mitch Ratcliffe
  • How do I find you?

    The problem I see is how will someone like me
    find you?

    Most of the new writers I have found had a book
    on the warehouse clearance shelves in big box
    retailers. Their $27.50 book was on sale for $5.98
    so I gave them a try.

    If I like them I buy other books they have written.
    On more than one occasion I have started a
    discounted book only to discover that (1) it is part
    of a series and (2) I like that writer. That leaves
    me setting that book aside while I buy and read
    that writer from book one on.

    This way of building a library simply isn't possible
    unless I can go to B&N and a sales card that allows
    for a $5.00 download.

    The other problem is that I pass the better books
    onto family & friends, giving that single book
    multiple readers. The ones I don't want to keep
    for re-reading in a year or two goes to the local
    VA clinic where the VA Bus takes them to the
    regional VA hospital. End result (excluding the tax
    deduction) is that a lot of people end up reading
    that single book.

    I can see your thinking working in a lot of non-
    fiction areas, but believe that the Dead Tree Book
    Society will live on.
    • Agreed -- the text has to be a gateway to the creator....

      and the other people involved with it, so that you can find me (an author)
      or find the editor who found me and see what else she has to suggest to
      you. There are a lot of new ways to link to value lying fallow, and the
      inability to pass along a book because of DRM, whether as a
      recommendation or as a contribution to the local hospital, is probably
      the most short-sighted design feature in the ebook today. Pass-along
      rates are important to growing a market.

      I think fiction can do this, too, btw.
      Mitch Ratcliffe
    • The finding needs to be spread around

      You describe a discovery process deeply linked to the book distribution
      process. In an ambient discovery environment, where comments,
      discussions and other reader-generated data is shared, discovery could
      be a lot easier, including in fiction. For example, Joyce's references could
      be linked to their sources, allowing you to find new authors through that
      work of fiction. Another alternative would be in detective fiction, where a
      vernacular term is linked to other uses in other novels of the same
      genre. In short, though, this is already the idea behind the book-sharing
      sites, such as I'm suggesting it should melt into the
      book, and that's just a start.
      Mitch Ratcliffe
  • Four blog posts in a month

  • eBooks in XML/URI, not crypto permits linkage

    Cryptography does not facilitate linking, bookmarking, creating overlays, annotations or references.

    Markup (tagging) and addressing does that.

    If eBooks were stored in an XML master copy, the tags (chapter, page, anchors, citations, images, etc., etc.) could be discovered, parsed and linked to from anywhere.
    All you need is a universal addressing scheme (URI references) to link to anything. This is what Tim Burners Lee's Semantic Web and RDF is all about.

    XML can be rendered to any reader, browser, whatever.
    The font size does not matter, as the tags are still there in the correct location, "hidden" for linkage software to discover. It can be linked to from any software that can resolve the URI reference (what a browser does when it follows a link to somewhere in the middle of a page using a URL). So a bookmark can point to anywhere (properly tagged) within an XML document.

    Tools similar to Google and Twitter and future browsers could gather an original ebook and the annotations (with URI links) of others together for presentation in a user-interface.

    Crypto is mostly useful for authentication, security, privacy and DRM. It is a distraction from the issues you are concerned with.

    This comment is a "gloss" on your article, it is linked to it in a very basic way. However, your article is not available in XML with tags on paragraphs, so that I can't link my comments to specific locations in your article and that might appear like tool tips when hovering over that section. Lots of cool user-interface issues here...

    I'd happy to keep my bookmarks locally and back them up somewhere like a future RDF/URI/XML-based
    • You're missing my point

      I am describing an environment of *selective* sharing and marking of
      books for different kinds of discover. Storing a book in XML, allowing
      tagging and markup would certainly make some of this idea work, but
      only within one copy of a book (or else, through a service that
      somehow propagates links and other markups to all copies of a
      book). The problem though with just Semantic Web technology, is that
      reading isn't, or need not be, a public act. Reading can be dangerous,
      too. Cryptography should be integrated into books, so that owners
      can track what they do (simple scenarios, like turning to the last page
      read for three different readers on the same device, a scenario easily
      imagined in a school), and share their thoughts in small or large
      circles, not just as a binary share/don't share choice.

      What about simply turning a digital copy into a family heirloom by
      having each owner sign the book digitally, so that a record of who
      read it could be displayed? Then, what if you only wanted to read
      Grandpa's comments? In both cases cryptography is necessary. I
      suggest that it is also desirable for books, because some ideas, some
      notes, some books themselves may be hidden by a reader for years
      before they share it with anyone. Consider someone who wanted to
      annotate a book their society considered dirty, and only wanted it to
      be revealed after their death? Cryptography, not RDF, not XML, or
      hyperlinking solves the problem. We need to be allowed to choose not
      to share our reading selves, too.

      The Semantic Web is an evolving vision, as is my description of an
      ebook, not an engineering spec. There are a ton of interesting UI/UX
      questions, which makes all these issues worthwhile to riff on.
      Mitch Ratcliffe
  • BookGlutton

    BookGlutton has some nice features, but the format and the
    community communication isn't yet near the kind of book I'm
    describing. When a book/document can be annotated,
    commented on, argued over regardless of the device it is
    read on and the context of its use, then we'll have something
    Mitch Ratcliffe