On Demand Books has installed an Espresso Book Machine in the New York Public Library’s Science, Industry, and Business Library and is offering to print books for anyone who comes by to request one. Their machine has been running since June (and will run until the end of the month), but the Internet seems to have only taken just notice of it and there was a flurry of publicity this past week. I went over to 34th Street to take a look at it on Wednesday afternoon (just after the New York Times visited, I think).
They’ve installed the machine prominently on the first floor of the library. It’s about the size of a small car and it looks like a bunch of laser printers were smashed together and a computer was stuck on top. Signs explain why it looks jerryrigged: this machine is a prototype, “On Demand Books Espresso Book Machine Model 1.5,” although the Model 2, about half the size and looking much more sleek, is on the way. While the press release suggested that anyone could come up and start printing out books, in reality the machine was cordoned off from the public and being run by an operator.
For this demonstration, there’s a list of 20 available titles: the usual assortment of out of print Open Content Alliance books (Dickens, Tom Sawyer, Beatrix Potter), a couple of scientific papers (Einstein, also out of print; a paper from the AMS), and two recent ones related to the venture: Jason Epstein’s Book Business, which made the case for machines like this being the future of bookselling in 2001, and Chris Anderson’s The Long Tail. And one odd one: Faulkner’s Three Famous Short Novels, published by Vintage in 1942, which is still in copyright (and in print). The operator suggested that Moby-Dick would take too long to print – because it’s so long, it would be printed in two volumes – and tried to get me to choose The Long Tail, which is nice and short. I wanted something that I’d actually read and I was curious about how the Faulkner volume wound up in the list, so I went for it. Was I sure I didn’t want The Long Tail? I was sure.
The operator clicked a button on the computer’s display and the machine soon started making printing sounds. This continued for the next fifteen minutes. First the pages of the book were printed; they were printed on standard 8.5” x 11” paper, double-sided. The Faulkner book has around 160 leaves; this took a long time, and was exactly as exciting as waiting for a printer to print 160 pages. When all the pages were printed, they were apparently moved to another part of the machine where glue was applied to one edge. (While the machine has translucent sides, it’s hard to see what’s going on inside it for the most part.) They were moved down to another printer, this one color, which printed the cover on thicker stock. The cover was then glued to the pages and folder around them. Finally, the book was moved to the last section, where it was clamped down and rotated three times to cut off the extra trim, making a book that’s about 5” x 7”. The waste paper dropped down to a bin at the bottom of the machine; the newly minted book came out a slot in the front of the machine. The operator picked it up and handed it to me.
How does it look? It looks like a cheap paperback. My copy wasn’t quite cut right and there’s a little spur of excess paper rising from the top right corner, which gives it a modicum of uniqueness. Like the other Open Content-printed books that I’ve seen, the print isn’t wonderful: they seem to be working from screen-resolution scans of the books, and they appear notably grainy when printed. It looks very much like a book that someone photocopied on a copier with the contrast set a bit too high. But like a photocopied book, it’s certainly legible. It’s worth pointing out that this grainy quality is a function of the scan rather than the machine: a copy of The Long Tail looked just like a PDF printed through a regular laser printer would look. It still doesn’t look like a regularly printed book, but it certainly works as proof of concept.
More importantly, what does it mean? While there’s certainly work that needs to be done on these machines, they certainly seem viable. Epstein proposed these machines as a solution for a single problem: the unavailable backlist. It’s not hard to imagine, however, that a decade from now the entire bookstore will have been replaced by one of these machines at the FedExKinkosBarnes&Noble. Holding my copy of Faulkner in my hands, the overwhelming feeling was one of cheapness: the book had been reduced, finally, to being a disposable consumer object, available as easily as a latte at Starbuck’s. The books that the Espresso was putting out every twenty minutes existed for demonstration purposes: although passersby oohed and ahed at the possibility of the machine and happily took the sample books, I sensed that the books probably wouldn’t be read.
We’ve noted here how young people don’t tend to keep CDs: when they buy them, they immediately rip them into the computer, often throwing away the packaging and the CD itself. Over the past five years, music stores have been closing at a precipitous clip; so have video rental stores. There hasn’t been a tremendous outcry about this: we get enough out of the convenience of the iTunes store or Netflix that we don’t care that Tower Records went under and that Blockbuster is struggling. What happens if the book goes in this direction? It’s certainly technically possible – both Google’s book-scanning project and the Espresso machine demonstrate that. But technology has moved faster than our sense of how our culture will be affected. There’s a discussion here that needs to happen.
if:book
A Project of the Institute for the Future of the Book
A footnote, which might not be read, but which might explain the tone I take in the post above.
Two stories about printing came to mind while I was waiting for my book. The first was Kafka’s “In the Penal Colony”, where the officer of the penal colony explains to an explorer the colony’s barbaric machine for execution:
The officer’s machine, it soon comes out, is a printing press that kills the condemned by tattooing their sentence onto their body over and over until death comes. Though the parts are all there in the Espresso, the comparison is a touch melodramatic.
The second was “The Paradise of Bachelors and the Tartarus of Maids,” Melville’s fable of the age of mechanical reproduction. Nobody bothers to read this story, which is a shame as it’s astonishing and worthy of Kafka. It consists of two parts: in the first, the narrator visits a lawyer’s club in London; he is astonished at what a nice time the men are having. The second half seems unconnected: the narrative moves from England to New England, where the narrator, who it turns out has a prosperous business selling packets of seeds, visits a papermill. The papermill unexpectedly is a Blakean hell, transforming the blood and sweat of the unmarried women who work there into product:
Melville never really explains what the link is between the two halves of the story. The reader is left to decipher how the bachelors’ life of pleasure requires the suffering of the women: to oversimplify Melville’s complex systematics, lawyers require paper, and somebody needs to make the paper. It’s a searing indictment of capitalism. But it’s also a plea for the reader to notice the complexity of the world: nothing runs unmixed, and it’s worth examining the world we live in. It’s a text that bears thinking about with relation to the Espresso: while it would be going too far to declare the Espresso will make people’s lives a misery, it will have unexpected consequences, and we do need to think about them.
There’s a neat visual pun in Melville’s story when you’re reading it in a printed book: as you read the pages on which are inscribed the story of the Bachelors’ Paradise, those pages are above the second section of the story, the Tartarus of Maids. The demanding work of the papermill underlies the bachelors’ pleasure (and, of course, under the pieces of paper that the words are printed on. The act of reading is going deeper into the physical book; it’s not a device that can be used in the screen environment, though perhaps putting one text below another in a scrolling field replicates some of it.
There’s an interesting article by Reinaldo Laddaga in the March 2007 issue of PMLA (v. 122 #2) that raises some significant points about the “culture” of reading and writing and how that culture is changing in recent years (but not without intense anxiety): “From Work to Conversation: Writing and Citizenship in a Global Age.”
The technology is moving faster than our cultural institutions (and our brains)!
Hmmm… I’ve been thinking about that handy little machine v. bookstore ever since reading Epstein’s book and wondered about how many people know what they’re looking for or can browse so effectively online that they can identify the books they want. Just recently a public library caused some controversy because they ditched Dewey for a B&N-style shelving arrangement precisely because they believe most discovery is through browsing the shelves rather than looking for a particular title. So if they’re right and browsing physical shelves is the preferred route of discovery over known-item searching, our technological future will need to account for that somehow. (Apparently browsing online has not supplanted physical browsing in this library.)
Thanks for the report from the front, by the way. I had read about this but knowing how it actually works and what comes out is much more helpful than a press release.
This machine is a fantastic concept and seems to be one future concept of book distribution. Although we might not be as cool and as instantaneous as this machine, BookSwim sends out requested books to our online book rental members through the mail, then send them back when finished, with no late fees, unlimited rentals per month, and free shipping both ways. BookSwim Book Rental Library Club can be found at http://www.bookswim.com.
I myself would love to see this machine in action. The real future would be to have people be able to print these at home!!
“In 2006 Americans purchased 575.1 million copies of mass market paperbacks. In 2006 Americans purchased 418.2 million copies of adult trade paperbacks. For those who are interested, 583.5 million mass market paperbacks and 420.6 million adult paperbacks will be sold in the U.S. in 2007. Those numbers will grow to 601.7 million mass market paperbacks and 429.3 million adult trade paperbacks by 2011.” Book Industries Study Group
This information begins to illustrate how print book reading grows as it diminishes. Many genres of publication are migrating from print including book format genres such as travel guides and schedule directories, phone books and encylopedia. But more than compensation is provided by the growth of monographic and fictional book works, the classical book genres, where the physical format retains exclusive economic, haptic and archival attributes.
Another momentum is the revolutionalry digitization of the print book production industry. Nowhere is this more apparent than in the migration of print production from off-set to electrostatic, from printing presses to high speed copiers.
Interfacing of the industry to the customer is also being digitized and not just by enhancing dewey decimals with plain signage. The comprehesive imaging of print books on screen has resulted in revolutionary new bibliographic access to books that can be delivered to paper (via re-imaging). This print-on-demand scenario has only begun its sweep across the prospects for the print book in the context of screen reading.
it occurs to me that browsing in a bookstore is
grossly inefficient compared to online browsing,
since an online environment can have _all_ books.
(not to mention the potential of doing a _much_
better job of saying why one book is like another.)
anyway, dan, even after having read the “footnote”,
i’m not really sure why your post has the tone it does.
i think the espresso machine will be _liberating_…
although i think it could be boiled down to something
_much_ less unwieldy, as evidenced by the “bookmobile”
the internet archive has sent all around the globe…
-bowerbird
gary said:
> where the physical format retains exclusive
> economic, haptic and archival attributes.
might watch your use of “exclusive”, gary… :+)
other than that, i agree entirely with you that
on-screen usage will drive print-on-demand…
-bowerbird
it occurs to me that browsing in a bookstore is grossly inefficient compared to online browsing
Questionable. It’s really an issue of the interface, I think. Amazon.com has done a great job cross-referencing stuff through their interface with features like Listmania! and “Customers who bought this book also purchased:”. But this is because they need to catch up to the physical interface of the bookshelf, which is more sophisticated a device for information display than people often give it credit for.
“efficient” is the superlative that bowerbird has chosen… a tricky one. If efficiency is simply speed, I’d quibble– I can browse, say, 50 poetry books on a shelf at Barnes and Noble pretty much at a glance, digging deeper when something catches my eye: a lot faster than I can click through 50 pages at Amazon.
But efficiency isn’t simply speed. There’s a signal/noise component built into “efficient” as well. 49 of those 50 poetry books speedily browsed on the shelf may not be to my taste (not unlikely if I’m browsing the limited stock at a Barnes and Noble). The more thoughtfully directed Amazon links may do a better job: instead of having to click through 50, I may need to click through two.
More efficient? In that way, yes. But sifting through the “noise” (on a shelf) provides multiple layers of meta-information (efficiently) in a way that the 2-click Amazon experience does not. So I’d say the “more efficient” title is still up for grabs.
I would like to circle back to “exclusive” attributes of print vs. screen. I did intend to identify exclusive attributes of print; legibility or immediacy of meaning, haptic efficiencies and simple persistencies, for example, that are and will remain exclusive to print. I don’t recall screen advocates being reticent to claim exclusive attributes for digital reading including attributes of automated discovery, multiplexed presentation and mediated currency, for example.
The exclusivity issue is important here. If the book will merge or reassign print and screen attributes as fungible relationships we should see this emerging; screen presentation with print attributes and print presentation with screen attributes moving toward an eclipsing reading mode. I suggest instead that we are watching print and screen move off separately, both with the momentum of digital production, distribution and delivery but each with exclusive attributes of readbility.
j. bushnell said:
> I can browse, say, 50 poetry books on a shelf
> at Barnes and Noble pretty much at a glance,
> digging deeper when something catches my eye:
> a lot faster than I can click through
> 50 pages at Amazon.
but you can only browse the books on that shelf,
not the 50 million others _not_ on that shelf…
> But efficiency isn’t simply speed.
good thing, because if it was, an online/offline
comparison would be completely laughable…
> There’s a signal/noise component built into
> “efficient” as well. 49 of those 50 poetry
> books speedily browsed on the shelf may not
> be to my taste (not unlikely if I’m browsing
> the limited stock at a Barnes and Noble).
> The more thoughtfully directed Amazon links
> may do a better job: instead of having to
> click through 50, I may need to click
> through two.
if you’re “clicking through” any books at all,
the system you’re using will not be “efficient”.
collaborative filtering will deliver books to you
using your ratings of various books in the past.
-bowerbird
collaborative filtering will deliver books to you
using your ratings of various books in the past.
With a 100% hit rate every time? Amazon currently does OK with this, but there’s definitely room for improvement, and unless it abandons the one-book-per-page system it’s been using, or hits the 100% perfect suggestion mark (to my mind an eternally receding horizon) there will remain a clickthrough issue. It’s a user operation inherent to the interface.
j. bushnell said:
> With a 100% hit rate every time?
before you were happy with those 50 books which
barnes and noble happened to have on the shelf.
now you want a perfection on 50 million books?
i said that bookstore browsing is “inefficient”
because the books on the shelves are limited,
and selected for their commercial viability…
unless your needs run in that specific direction,
you ain’t gonna be able to find what you want…
of course, you will not know what you’re missing,
not until you are presented with a better system.
> Amazon currently does OK with this
amazon has collaborative filtering that only a
capitalist could love, since it’s based on the
behavior of _purchasing_ books. when we have
a system based on our _ratings_ of each book,
you will see how powerful the mechanism can be.
-bowerbird
before you were happy with those 50 books which barnes and noble happened to have on the shelf.
now you want a perfection on 50 million books?
Well, I actually said that it was pretty likely that the 50 Barnes and Noble books would not be to my taste. But I’ve never really been talking about which is going to make me “happy” or which I “want.” I’m trying to stick with the question of which is more efficient.
It’s tough to imagine a material interface that will allow one to browse 50 million books efficiently. I’ll agree that the collaborative filtering systems we have at present (we can look at the Recommender at LibraryThing if you want to get away from the capitalist spectre of Amazon) are powerful mechanisms, but (unless they someday get to that elusive 100% hit-rate) they don’t completely eliminate the need for browsing, and browsing online (at present) is slower and more cumbersome than scanning a bookshelf.
I don’t disagree that online browsing can yield vastly better results, because of the obvious advantages (not having to be limited by shelf space, finite stock, etc). We could certainly say that this makes it more efficient, along one axis. But it also cuts out certain levels of metadata that a physical shelf in a bookstore or library retains and (efficiently) delivers.
jeremy said:
> browsing online (at present) is slower and
> more cumbersome than scanning a bookshelf
when the book you want to browse is on the shelf,
i agree. odds are, however, that it will not be.
even if it is, we’d have to define what purpose
the “browsing” is meant to serve. but i suppose
we’ve reached the limit of useful dialog here…
-bowerbird
(this would be easier in CommentPress format)
(there is always a friction death of discussion here just because the threads meander…but not because the topics fade. Discussion of relations of automated and physical searching are important to evaluation of print and screen research and methods of their combined use) (Just such an exchange has germinated on SHARP-L.)