New! Espresso Book Machine 2.0


Anyone who has seen the Espresso Book Machine 2.0 designed by On Demand Books is likely to be impressed by its efficiency in producing a printed book from one of more than 2 million public domain digital books offered by Google Books. At the Harvard Book Store in Cambridge, a terminal at the store’s entrance allows one to browse titles. Once a book is identified as available, an assistant reviews the electronic images to make sure that the book has been adequately filmed.

The next step is to walk to the back of the store, where the “Espresso Book Machine” (or, as the Harvard Book Store calls it, “Paige M. Gutenborg”) abides. This looks like a fancy photocopier with binding accessories (see below). The assistant types a number into the machine; the machine does the rest. A cover is produced in color and lowered face-side down onto a binding table. Copies of pages (recto and verso) shoot onto a tray above the binding table. Once the pages are printed, they are clamped, producing a bookblock, which is turned vertically, spine downward, so that the spine can be milled. A glue bar applies a coat of glue to the spine. The pages are then lowered onto the cover, which is lifted flush against the bookblock and clamped in order to allow the glue to set. The “book” is then cropped on three sides, and sent down a chute. It arrives still warm from being photocopied and smelling of glue. The paper quality is acid-free and superior to the paper most photocopying machines provide. The entire printing process for a 400-page book takes about 8 minutes. The books cost $8.

Versions of these machines exist, have been exhibited, or are coming to about 26 locations in North America, Australia, Egypt, and the U.K. Their most obvious use is in libraries. Instead of photocopying (and damaging) books, one can now use this machine to produce far better working copies of texts than most photocopiers provide. The machine costs $100,000, which will limit its purchase by university libraries undergoing severe budget cuts. Nevertheless, its potential is promising.

The press release regarding Google Books’ agreement with On Demand Books can be accessed by clicking here. See also On Demand Books’ web site by clicking here.

The following YouTube video is offered by On Demand Books:


22 Responses to “New! Espresso Book Machine 2.0”

  1. Eleanor Shevlin Says:

    The University of Michigan has also purchased an Espresso Book Machine. Evidently, at this time the titles available are from its digital collections and serve as samples (that is, UM does not yet seem to be selling books). This choice would seem to encourage the digitization of a library’s holdings while also facilitating their dissemination.

    UM’s plan for the machine also demonstrate its versatile potential:

    What is the Library planning to do with it?
    There are many possible services that we can provide through the EBM. Initially, we’ll be able to begin to provide affordable, high-quality reprints of out-of-copyright books from the Library’s collection and from resources like the Internet Archive. Over time, we anticipate offering other services such as:

    Small runs of printed books produced by classes, such as anthologies of creative writing
    Printed copies of proceedings of University conferences and events
    Printing and binding course materials
    Self-publishing for Ann Arbor authors

    This list of possibilities indicates that this machine would serve courses, the community, and professional, scholarly activities.

    Here are some more details from its FAQs page:

    Q: Are you printing Google books?
    A: Yes, we are printing public domain books from several sources. This includes books from the Google digitization project, as well as items from our own digitization efforts such as the Making of America collection and from the Internet Archive’s Open Content Alliance. Google-scanned titles are also available via the EspressNet catalog, courtesy of On Demand Books’ recent agreement with Google.

    Q: Will you be selling in-copyright titles via the EspressNet catalog?
    A: We are considering the implications of selling the in-copyright titles. For now, we are limiting sales to the public domain items available on the networked catalog.

    Q: How long does it take to fill an order?
    A: If a book is already prepped and loaded to the catalog, it can be printed in 7-10 minutes. If interior and cover PDFs need to be created for new titles, we ask that customers allow 2-3 business days for fulfillment.

    The comment that “The machine costs $100,000, which will limit its purchase by university libraries undergoing severe budget cuts” from the Espresso promotional material is a thought-provoking one…and signals another example of a growing digital divide.


  2. Anna Battigelli Says:

    The U of M site concedes that the EBM is not a profit-driven enterprise. Because the EBM does protect books, however, it seems well worth the expense.

    And, yes, it is an example of the ongoing digital divide.


  3. Eleanor Shevlin Says:

    Yes, the purchase price does seem worth it. In fact, I suspect many regular copiers in fairly large academic departments cost 12,000 to 18,000 (some depts. might rent though), so the price actually seems reasonable to me.

    The various possible uses for the EBM that Michigan presents also seem quite thoughtful. Planning, for instance, to offer Michigan authors a chance to self-publish seems to be thinking outside the box and provides a potential service to the community beyond campus.


  4. Anna Battigelli Says:

    The benefits I see are the following:

    1. Allowing researchers to produce nicely bound working texts
    2. Providing access to out-of-print books
    3. Facilitating course creativity by enabling instructors to create coursepacks and anthologies
    4. Allowing academic journals and academic texts to exist online with on-demand publication
    5. Making self-publication possible

    I would be interested in hearing from book store owners whether they feel that purchasing the ebm is a commercially viable enterprise. I would also be interested in knowing how members of the publishing world (both commercial and academic) and how librarians feel about this innovation. AB


  5. Eleanor Shevlin Says:

    On the American Booksellers Association website, a September 30th piece, Espresso Book Machine 2.0 Makes an Impressive Debut, discusses the use of ebm in several independent bookstores in Vermont, Michigan, and Washington state.

    The Northshire bookstore in Vermont, which had acquired its ebm last February with the immediate goal of offering “more choice for our customers,” has found that the prime use so far has been for self-publishing.

    Other bookstores also seemed to feel that self-publishing purposes would be the most popular, initial use of ebm. One owner noted that “POD was becoming an important part of the book industry, and would only become more important with time.”

    As for commercial viability, it seemed to early to assess the bottom line, but most seemed to think it was a good investment.

    It has often been said that the new digital age, specifically the Internet, has parallels with manuscript culture in its ability to make anyone an author. These bookstore owners see POD as a strategic response to the Internet, and it seems as the emb offers another case of ongoing remediation of print as it interacts with the digital.


  6. Anna Battigelli Says:

    It’s interesting to see how book stores define themselves according to the uses to which they put the ebm. Northshire Books (Manchester Center, VT) focuses on self-publishing. They offer OCR scanning of texts, editing, and so forth, making themselves a small publishing venture. By contrast, the Harvard Book Store features re-publishing (downloading public domain texts from Google), thereby emphasizing the ebm’s scholarly (as opposed to commercial) value. (Self-publishing is also an option, but it’s not featured). To turn to a library, the University of Michigan has not yet offered self-publishing, perhaps hoping to emphasize the ebm’s scholarly function or its value for preservation.


  7. Eleanor Shevlin Says:

    And/or perhaps how patrons of various bookstores/libraries define the bookstores and libraries through the services they use. Among the possible uses for ebm, the Harvard bookstore’s ebm webpage notes,

  8. Provides authors with affordable, flexible printing options. There are no minimums, and you retain full rights and complete control of your work.
  9. Looks forward to printing your novel, personal cookbook, family genealogy, memoir, dissertation, personalized gift, and more.
  10. The UMichigan library lists similar possbilities. In the case of Norshire bookstore, it seemed that its customers have promoted the self-publishing use over others.

    More and more university bookstores are now being run by commercial entities (U of Penn’s, for example, is a Barnes & Noble enterprise, merging commercial with the scholarly. Their customers are almost certainly overwhelmingly students and faculty, however, so their needs and interests will dominate. On a somewhat related note, Barnes & Noble is apparently a leader in commercial ownership of college bookstores, and its position in this market would seem to accord it a decided advantage in promoting its e-reader, Nook.


  11. Anna Battigelli Says:

    Good point about users defining stores and the uses to which they put technology.


  12. Benjamin Pauley Says:

    Something I find encouraging about this announcement is that it suggests that some libraries are beginning to move on the possibilities afforded by their having received copies of scans performed through their partnerships with Google. There’s been a lot of concern that Google was poised for an effective monopoly in digitized books. These kinds of print-on-demand projects, though not directly parallel to the Google Books service, at least suggest that Google’s partners aren’t just sitting still.

    I do wonder how the ebm arrangement will compare to another recently-announced partnership between Michigan and HP to provide print-on-demand books. (HP’s press release is here.) The service is called BookPrep, and is conspicuously labeled “Beta.” I submitted a non-binding preorder for a title to try to get a sense of the pricing, but haven’t heard anything back after a bit more than a month. Which I suppose isn’t too promising, come to think of it…

    It’s actually quite interesting to me to see the various incarnations that these scans are taking. Books originally scanned by Google at partner libraries are available, by my count so far: through Google Books itself (with some libraries providing links from their catalogues to the scanned texts—a good way, incidentally, of confirming the identity of the books that Google scanned); through HathiTrust; through print-on-demand services like BookPrep and now ebm; and through a partnership with Barnes & Noble.

    The Barnes & Noble deal is an instructive cautionary tale, though. You can download a text originally scanned by Google for use with Barnes & Noble’s eReader software and with the Nook (which we now see was probably the point all along, though Barnes & Noble’s deal with Google came before that hardware was announced). Rather than packaging the page images, however, the Barnes & Noble eReader store presents you with the (frequently very bad) OCR’ed plain text. This move makes sense when you think about the need to have text that can be re-sized and re-flowed for use on an e-reader, but of course its consequences for handpress-era books are utterly perverse. Barnes & Noble repackages Google’s vanilla .pdfs in a protected .pdb format, so you’d have to download their free eReader software to see what I mean, but Barnes & Noble’s version of, say, Defoe’s Augusta Triumphans is, if possible, worse than useless.

    I trust that the ebm arrangement won’t make the same mistake, as those texts are destined to be printed, rather than read on a screen. (The previews at the BookPrep site are certainly page images, so that seems like a promising precedent.

    While I find the proliferation of outlets for these scanned texts to be promising (a false step like the Barnes & Noble arrangement notwithstanding), I find it troubling, as well, given the sketchy state of bibliographical data at Google Books. Google doesn’t really describe their holdings in terms that are useful in cases where bibliographical detail matters, and that confusion seems likely only to to intensify as the poorly-identified surrogates are propagated into new settings. (Perhaps this won’t be so problematic in cases where it’s actually the partner libraries who are doing the re-packaging of the scans that Google performed.) Google does seem to be interested in addressing the metadata problems, but I wonder whether any fixes they make will catch up to the documents that have been disseminated through other channels.


  13. Benjamin Pauley Says:

    [Quick note after actually having a look at the Harvard Book Store’s site.] It’s interesting to see that, for eighteenth-century texts from Google Books, the site is presenting Google’s id code in place of an ISBN. So, for instance, the first copy of Defoe’s Farther Adventures of Robinson Crusoe that appears for this search is said to be “JMoNAAAAQAAJ.” A quick check of my database at Eighteenth-Century Book Tracker shows that that’s a copy of the first edition (ESTC N4783) from the Bodleian. Right now, Harvard Book Store’s search system doesn’t allow for any kind of complex search that I can see, but this could be an interesting development.

    If Google’s codes are going to become de facto identifiers for copies of texts that they scanned, it makes me think I may need to make it so visitors to Eighteenth-Century Book Tracker can search the database “in reverse,” as it were, starting with a link or Google ID that they need to identify.


    • Anna Battigelli Says:

      Is the replacement of a Google id for an ISBN a problem in that the ISBN might more reliably lead to complete and accurate bibliographical information than the Google number? Looking through the HBS’s site, I found the bibliographical entries scanty. This highlights the value of Eighteenth-Century Book Tracker. I’m interested in the reverse search function.


      • Benjamin Pauley Says:

        I don’t think I’d say that the use of a Google id where there’d otherwise be an ISBN is a problem, necessarily—after all, it’s not like there is an ISBN to be had for these books, so using a Google-generated id is filling a void and providing at least some way to identify the book. I have a reverse search function set up for my own administrative purposes, but I’ll put some thought into how best to make it available.

        I think Eleanor’s hypothetical is an interesting one: if this were in a library rather than a book store, might they have opted for some other identifier, like an ESTC number or an OCLC entry number? My not-entirely-well-formed hunch is that an OCLC number would have been the likelier choice if entries for an Espresso catalogue were being generated from a library’s system, but those have their own problems where texts from our period are concerned.

        And I think you’re right, Anna, that the HBS entries are scanty on bibliographical information. It’s actually a very curious collision of worlds, I think: it’s set up like a book store, but with the reach of a library (or, rather, of a somewhat indiscriminate assemblage of books drawn from lots of libraries). You see all of the texts from Google Books alongside those (to my mind) dubious “reprint” editions that you also see at Amazon—ones that frequently just print and bind text nabbed from Project Gutenberg. So there are tons and tons of books to be had (I liked the quote from the Blackwell’s exec Eleanor pasted below), but it can be pretty tough to know the provenance of most of them.


      • Anna Battigelli Says:

        I should have been more careful in my previous note. For the purposes of searching modern texts in the Harvard Book Store’s database, having access to both ISBN and Google numbers, when both are available, might be helpful. For earlier texts, OCLC and ESTC numbers (and Google #s, where available) would be helpful, however problematic those databases are.


  14. Eleanor Shevlin Says:

    Many thanks for these very informative comments, Ben. I was not aware of the HP BookPrep project. But I did view the U of Michigan Library‘s acquistion of this machine as an extension of all its work in digitization projects and, in part, as a step toward making use of its work with Google. That it was Harvard’s bookstore and not its libraries that first purchased emb is also potentially interesting in terms of Harvard’s approach to the changing landscape. I wonder if emb was housed in one of Harvard’s libraries if books such as Defoe’s Augusta Triumphans would be identified solely by the Google ID or if an effort would have been made to provide more bibliographical information.

    Your worries that the bibliographic problems affecting so many of Google Books now have even more opportunity to spread and multiply faulty or erroneous information seem to be worthy concerns. I might add that the opportunity for spreading misinformation and the like is not limited to Google. Instead, many proprietary databases have that same potential for bibliographic errors, though to a lesser degree.

    Although in many ways quite different from POD, Dover Publication‘s longstanding role in making out-of-copyright works very cheaply available comes to mind. These works also often have bibliographic issues. It is interesting to see that Dover was acquired a few years ago by Courier Corporation, a company very much involved book manufacturing, fulfillment, and the like.


  15. Eleanor Shevlin Says:

    I left out a few comments I intended to include in my previous response.

    One is the link to HathiTrust, a recent digitization enterprise that we have not discussed (though it deserves mention and should have been noted in the remarks about feasibility issues related to Darnton’s suggested solutions.)

    Also, and I think Matt Wilkins made a similar point in our discussions about the Google Book Settlement, I don’t think that Google ever intended to become involved with the bookselling business but has now found itself in that position (and probably making the most of it) because of the lawsuits. It is thus interesting to see how Michigan is at once very much cooperating with Google, while also maintaining its own independent courses. Michigan, we should remember, was a leader in microfilm collections.

    And, finally, Ben, the merits and contributions of your Eighteenth-Century Book Tracker project grow with each of these new developments. I wonder if you have discussed any aspects of the “rights” to your work with the university. It would be an injustice if all of a sudden developments arose that made your institution want to claim some sort of right to your work because it’s hosted on a university server, etc.


  16. Eleanor Shevlin Says:

    As I was looking for an unrelated dcument in my spring 09 Word folder tonight, I discovered a folder labeled “Espresso Book machine.” It contained some material that appeared this past April when ebm made its debut in Blackwell’s bookstore at Charring Cross, London. I had forgotten I had even started the folder.

    One item was a brief article from The Guardian‘s. A few things are interesting in this piece. For one, Blackwell paid $175,000 for its emb. In the interview, Blackwell emphasized the wide promise of ebm for bookselling as a way of leveling the playing field:

    “This could change bookselling fundamentally,” said Blackwell chief executive Andrew Hutchings. “It’s giving the chance for smaller locations, independent booksellers, to have the opportunity to truly compete with big stock-holding shops and Amazon … I like to think of it as the revitalisation of the local bookshop industry. If you could walk into a local bookshop and have access to one million titles, that’s pretty compelling.”


    I am also supplying a DIMACast (Digital Imaging Marketing Association podcast) interview with Thor Sigvaldason, Chief Technology Officer of On Demand Books, from October 9, 2006. Here’s the link to the interview (the actual interview starts at 6.17 or so on the ticker). Besides offering historical perspective, it provides some fascinating information. The early machines, for instance, were made by hand in St. Louis, MO. At the time of the interview, the New York Public Library already had one installed as did the World Bank.

    I am also including a link to another DIMACast interview that took place this August 31, 2009 with Andrew Pate, Vice President of Business Development for On Demand Books as a follow-up to the interview with Sigvaldason.

    My discovery of these interviews came about because of a conversation that I had a long time ago but that I had not comnnected at all with emb until tonight. One day in the summer about six or seven years ago, one of my former condo neighbors,Thor Sigvaldason, mentioned his work on this print on demand machine. (I believe we were participating in a building “spruce-up” day.) He knew that I did book history work, so he thought I might be interested in a project he was undertaking with Jason Epstein. We proceeded to have this discussion about ebooks, future of publishing, the book, etc.. I now realize his project was the emb.


    • Anna Battigelli Says:

      The interview with Andrew Pate (8/09) was interesting, so thanks for that, Eleanor. I was particularly interested to learn that book stores and libraries set their own prices for the books. I was surprised (but perhaps should not have been) that half of the ebm’s users use the machine for self-publishing.


  17. Eleanor Shevlin Says:

    What I found interesting from comparing the two interviews is that self-publishing was obviously on the radar as a use very early on. So, too, was the aid it would offer independent bookstores. That the New York Public library and World Bank were the very first initial users is also worth some thought.

    That each owner of emb sets its own price underscores the ties to copying outfits (I wonder what, if any, effect emb will have on the “publishing” business done by Kinko’s and Staples).


  18. Benjamin Pauley Says:

    I decided to buy an Espresso Book Machine-produced text through the Harvard book store’s web site. The book arrived today, so I thought I’d offer a quick report.

    The sale through Harvard’s book store was painless (setting aside the difficulties of knowing just what book you were getting discussed above—if you know what you’re looking at, the actual transaction is perfectly smooth, with good email feedback along the way). Add shipping to the $8 cost brought the total up to around $13, which still seems like a bargain. (It’s about what lots of classroom editions cost, though of course without the editorial apparatus.)

    The book itself is pretty nondescript—a simple blue and white cover with information clearly drawn from Google’s metadata (the front cover has “Defoe, Daniel,” and the title on the spine is truncated: “The Farther Adventures of Robinson Crusoe: Being the S”). The size is comparable to a small Penguin or Oxford World’s Classics volume, with the same sort of glued binding. The book feels solid, but of course I wouldn’t expect it to stand up to the sort of sustained re-reading that the books I teach most often receive. The paper is comparatively heavy (about the same weight as that in my copy of Gaskell, to take an example lots of readers of this blog probably have to hand), and quite smooth—much smoother than, say, a Penguin. It’s easy to write on in pencil, for what it’s worth.

    I guess it shouldn’t be any surprise, but it’s worth repeating that this is nothing more and nothing less than a printing of the file from Google Books bound into codex form. That means that there haven’t been any extra steps taken to clean up the print—if the text on a page is hard to read on your screen, it’ll be equally hard to read in the printed book. Though there are generally usable margins, I also see some pages where the page number at the top of the page is running off the paper. If there are fold-out maps, etc., you will likely not see the whole thing.

    On the other hand, buyers of these books benefit from Google’s practice of just scanning everything. The very first page is boilerplate (“This is a reproduction of a library book that was digitized by Google…”), but from there out, you see everything that Google scanned: including the bookplate, a pasted-in bookseller’s catalogue description, a portion of the Bodleian shelfmark, an owner’s inscription, Bodleian stamps scattered throughout, and eleven pages of advertisements at the end of the book. At least in the instance I have in front of me, you see every opening as it appeared in the scanned book—no blank pages have been removed, for instance, as I understand is sometimes the case with texts at ECCO.

    So I suppose this is a rather too-lengthy way of saying, “Yep, it’s a paperback all right,” but that is, in fact what it is. And really, that’s not so mean a feat when you think about it. I can see these print on demand copies becoming very useful for students and scholars. At these prices, it would be feasible to buy copies of several different editions of a text for comparison, which would be a real boon to certain kinds of projects.

    There are lots of potential pitfalls here, of course, over and above the bibliographical problems we noted earlier. There are sometimes problems with Google’s scans (missing pages, etc.), and it would be very disappointing to purchase a book only to find that the text was incomplete. Then, too, Google has often scanned more than one copy of an edition at different libraries; it would be nice to know if one of those scans were better than another.

    Right now, I think that to make the best use of this service, you really have to double-check the integrity of the text at Google Books before clicking “Buy”—you can’t rely on the level of quality assurance that’s tacit in bookstore transactions. But with those cautions in mind, I think this really has potential. I’m going to go put my credit cards in a locked box now…


  19. Anna Battigelli Says:

    Hi Ben,

    Thanks for description of the ebm book. I also purchased a book and liked the feel of the page and the crispness of the image. Having watched the glue be applied, I did wonder whether such texts would withstand a semester’s stress. Still, I was impressed. This is a promising technology for teaching and for scholarship, especially if prices remain reasonable.


  20. Eleanor Shevlin Says:

    Yes, thanks so much for the description, Ben. This discussion and particularly your comments about the Google source for the Defoe work brought to mind Kessinger Publishing and other outfits that seem to be taking out-of-print works and either scanning copies they have or using Google digital copies to produce the copies they sell. (A review comment on by someone who purchased one of Pope’s works alerted me to Kessinger using Google digitized copies for “their” books; the reviewer noted that the pages actually had the scanned by Google “imprint” on the pages). I’ve mentioned this trend in earlier emob posts, but it seems worth mentioning again in this context.

    When these works show up on Google Book Search, often only limited previews are available (even though the work is, say, an 18th- or 19th-century edition that is out of copyright or, perhaps I should say, *was* before Kessinger or the like reproduced it). Indeed, sometimes only “snippet views” are available, and at other times no view at all, as is the case with this work. Adding to bibliographic chaos, often the source work is not acknowleged. By the way, did Google Book Search eliminate the “snippet view”? Interestingly, when I did a quick search for Kessinger and Pope, the list came up with three choices to the left: 1) Limited preview and full view 2) Full view only and 3) Public domain only; there was no snippet view listed. When I clicked on “Public domain only,” most of the Kessinger titles for Pope disappear. While I need to do a bit more investigating about Kessinger’s copyright claims (as well as the “snippet view”), the fact that many of the Pope titles disappear convey the impression that Kessinger is claiming copyright.

    I looked to see if Kessinger had a copy of Defoe’s Farther Adventures of Robinson Crusoe, and it does. Especially interesting is the price that these outfits are charging in comparison to the price you paid for your book, Ben. For the Farther Adventures, for example, Kessinger is charging $27.95.

    And here’s what the one customer who bought the Kessinger copy on Amazon had to say about it:

    I did not read this book,, It was printed in the old type printing hard to read, and I returned it. Had I known it was printed that way and not in regular printing, I would not have ordered it.


  21. NEH Digital Humanities Startup Grants: Funding the Future « Early Modern Online Bibliography Says:

    […] website that anyone can produce using a machine like the Espresso Book Machine (see an earlier EMOB post. An equally fascinating feature of this project is its dual display of English and Arabic text as […]


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: