An Active Reading Machine

Reading is simple enough, right?

• Annotation and Markup: When we read non-fiction material, we often find it useful to markup the text and write notes to ourselves. Paper systems combined with the respective writing device allow such annotation and markup quite easily. We may highlight text with a colored pen to denote importance, or a questionable passage. We may also underline text in a similar fashion using pen or pencil. Often most useful, however, is the ability to write comments one the paper itself. We may draw a line to a particular passage and then write about our understanding of it or some possible extension of it. We may make comments about possible problems with a statement or we may provide evidence of its truth. In any case, paper makes it very easy for us to customize and extend a document.
  
  Computer display systems seldom provide us with the ability to perform either of these tasks, much less both. Some systems allow the reader to highlight sections of text, but provide little in the way of annotation. Products like Microsoft Word allow the author or reader to highlight sections of text in special colors, but these are generally considered writing tools and display text in a writing-based format. Further, the steps necessary to "use" the highlighter are often buried in menus, making their use more of an annoyance that an aid. Other systems that do provide for annotation are often clumsy and inefficient. Sure, Adobe Acrobat allows you to place notes in the margins, but you have to move the mouse to the margin, click a spot, type in the text of your note and then issue a command to save the note. Once all that is done, there is some issue as to how Acrobat thinks it should display the note. Not quite as easy as just jotting a quick note in the margins with a pen.



• Instantaneous Paging: Often when reading, particularly when the subject is complicated or hard to follow, people flip back and forth between pages. By the time we have moved forward several pages, we sometimes need to refer back to previous statements or arguments. In paper books, this is relatively simple. We turn a few pages while holding our current place with another finger. We may even flip forward and back again and forward again, comparing what an author has stated previously to what he is stating now. While computer display systems certainly provide us with the ability to page back and forth, it is not nearly as easy to do the quick alternating paging that books allow. First, you can't stick your thumb into a computer document. You could create a bookmark at a certain point, but returning to that bookmark would most likely not be a simple matter. Further, many display systems, especially on machines with low memory, have trouble switching between one display and another. Especially when paging quickly back and forth, users often experience a lag in refresh of the screen.



• Simultaneous Displays: Often when reading advanced material, or when cross-referencing other documents, people will have open any number of books simultaneously. If I am reading a Russian article on computer viruses, I may have open the article itself, a Russian to English dictionary, a special Russian Technical glossary, and perhaps an English language reference on viruses in case I don't understand the technical methods or implications of the article. I can have each document simultaneously open to a given page and refer to any of them almost instantaneously. Computer display system try to accommodate such behavior using windowing systems, but this solution suffers from problems.
  
  First, with paper I can simply glance over at another document. On a computer display, I will most likely have to minimize some window or select some icon for full display. Also, as with rapid paging, quickly switching between several different windows often creates significant lag times. For instance, say I need to look up a Russian word with which I am unfamiliar. Indeed, say it has a rather strange spelling (as Russian words notoriously do). In the paper world, I can quickly glance back and forth between the word in the article and my dictionary until I find the entry I need. In a computer display, I can try to do something similar, but lag time will hinder me, or I can try to memorize the spelling of the word and then switch to the dictionary and seek it out, hopefully remembering the exact spelling while I search.

What's so bad about my desktop computer, really? (or, What's so great about paper, anyway?)

• Lack of mobility: People simply don't carry their desktops around with them. Instead, they sit in one place and wait for the user to return to them. The serious reader often wants to take a document elsewhere, or instance a bed, an airplane, or the bathroom. The desktop metaphor explicitly disallows this sort of behavior. Your display space is fixed to one location. Now, the desktop metaphor could allow for multiple desktops in different locations, but even then, the reader is limited to a finite number of display spaces.
  
  Similarly, one cannot hand their desktop to someone else for reference. I cannot walk down the hallway with my desktop, enter a co-worker's office, point to my desktop and say, "look at what this document says," or "can you clarify this for me." Paper, on the other hand does allow such physical transfer or reference of information. Admittedly, I could walk over to my co-worker's desktop and say, "pull up document so and so," then wait while we retrieve document so and so, then say, "OK, go to page so and so," and wait while we navigate through the document to that page, and then say, "look at what this document says." I could do that, but it would be much easier to walk over to her office and point to the page I already had displayed for myself.



• One size doesn't fit all: Not only is the desktop metaphor limiting in its physical immobility, it is limiting in its display size. Like it or not, a desktop comes in one size. By this I mean that a desktop's size cannot change based on the type of document being displayed. It is the same size whether displaying a post-it note sized memo, a paperback novel, a newspaper page, or a huge state map of Texas. Yes, desktops can be built in different sizes, but I will rarely know ahead of time that I need one single size. A different tactic would build the largest conceivable computer display and use it for all documents regardless of size. Unfortunately, the largest conceivable display would only be needed in rare cases. The rest of the time, our documents would be dwarfed by the empty screen surrounding it. Paper, however, makes it easy to have many different sizes of displays each suited to the document being used. If we only need to read a small note, it is usually written on a small piece of paper, but if we want to view a map of Texas, it is printed on a large, folded piece of paper that can either display the entire map or can be manipulated (or mangled as the case may be) to display only the relevant parts of the map.



• Humans weren't built for sitting: One of the most obvious features of the desktop metaphor is that behind your desktop is a chair. While thousands of dollars get spent trying to create a chair that's comfortable to sit in for long periods of time, the anatomical reality of the situation is that humans didn't evolve in an environment where sitting all the time was a useful trait. Maybe if their had been some environmental incentive for our primate ancestors to remain sedative, this wouldn't be a problem, but as plenty of back-pain sufferers can attest, it is. Indeed, chimpanzees and gorillas spend lots of time lounging around, laying in piles of soft leaves. Given a choice, I'm inclined to think most readers would prefer the chimp way of things over the desktop way of doing things. Unfortunately, a reader tied to the desktop metaphor doesn't get a choice. Either you sit in front of the desk and read, or you don't read. You could stand, but the desktop isn't really designed to read this way for very long. Paper clearly does not suffer this painful limitation. For the most part, you can read a paper document almost anywhere, in almost any position. You have to have a light source and you can't have something blocking your line of vision, but beyond that paper functions most everywhere[2].



• Paper as a human artifact: Paper has a long history, and has displayed almost all of mankind's greatest documents. As such, paper has become an intimate part of our cultural heritage. Why else do we carefully preserve paper documents such as the U.S. Constitution or the Declaration of Independence. There are surely enough copies by now that we don't really need the originals, but we keep them anyway. They are artifacts valuable not simply for the documents they display, but the part they played in our history and the formation of our society. Let's face it, people wouldn't line up at the national archives to see a computer display of the U.S. Constitution under glass. More than simply being cultural artifacts, however, paper documents are individual artifacts. Particularly in the case of books, individuals become attached to specific volumes or editions of a document. A fan of Melville, given a choice between the Penguin version of Moby Dick and an original first printing would take the first printing. Indeed, it would likely be a treasured item. Not because the document contained therein would differ in any substantive way from the Penguin version, but rather because the book as an artifact holds more meaning than the words inside it.

But paper has problems too.

• For instance, as noted above, paper has to have an external light source[3]. Computer display systems are emmisive, and therefore are their own light source.



• Paper documents are not dynamic. Once they have been printed, they do not change. Paper does allow us to annotate, which is sort of like changing the document, but computers allow us to change the original document itself, not just append to it. Further, computer display system can change in real time. To demonstrate how an engine works on paper requires several panels depicting changes over time. A computer display can simply show an animated engine diagram that the user can stop, start, or even reverse.



• Publishing in paper presents a barrier to entry. Some may argue this is a good thing considering the proliferation of useless or plainly wrong documents on the Internet, but the paper metaphor limits the power of the press to those who can afford one. Computers allow people to publish information quickly and easily, essentially dropping the price of the press considerably.



• While the desktop metaphor makes a single document unwieldy, the paper metaphor loses this advantage the more documents one has. When compared with a library of over 1000 books as well as numerous magazines, journals, and photocopied materials, it becomes relatively easy to lug around my desktop. As a collection, documents take up much more space and are less mobile than the computer.



• Finally, it can often times be extremely difficult to track down specific information in a paper document. Computer display systems often provide the ability to search for specified strings, greatly enhancing our ability to find the information we need from large documents. Indexes and tables of content attempt to satisfy our information seeking behavior in a paper based text, but none can compare to the ability of computers to perform full-text searching.

So what's this active reading machine, and why is it so great?

• If it looks like paper and feels like paper ... It probably is paper. The active reading machine of the future will be paper. In all of the above discussion of advantages and disadvantages of various models, the advantages that paper holds over computer based displays is intrinsic to the physical medium. Conversely, almost all the advantages of the computer based systems lies not in the physical medium of the computer, but in the power gained from digitizing information. Searching text is not specific to emmisive display systems connected to computers, it is only made easier by the computing power of the processor. The ability to fold paper, on the other hand, is a property of the physical media itself. To put it another way, we could recreate the former activity on paper (albeit much more slowly), but we could not recreate the latter on a computer screen.
  
  Given this fact, it only make sense that we try to recreate the advantages of computers in the physical medium of paper, rather than the other way around. Certainly this requires us to make drastic changes to our notions of what paper is, but it is possible. On the other hand, no matter how much you change the display properties of a monitor, you cannot change its physical characteristics so dramatically as to make it really resemble paper. What we really want to achieve is what Neal Stephenson calls "smart paper[5]."



• Smart paper is smart because it has electronic circuitry embedded into its fibers. Realistically this circuitry does not need to be very complicated. Each node of the circuitry controls a physical "pixel" that can be rotated to display differently colored faces. In the simplest instance, the pixel would be spherical and would have only two display states, black and white. By sending a particular pattern through the electronic circuit, a pattern of black and white pixels would display text[6]. It is reasonable to imagine that we might instead use cubical pixels and thereby gain four more display states. In any case, the maximum information that any pixel needs to have in this case is three bits. If these pixels are small enough, it is perfectly conceivable that smart paper could be two-sided just like normal (dare I say dumb) paper.



• We can write on smart paper too. By using a magnetic stylus, we can externally affect the orientation of pixels. Moving the stylus over the page leaves a trail of altered pixels, much as a standard writing device leaves a trail of graphite or ink. In this way we can add marginalia or highlighting to a smart paper document, by adding a few more bits of information to each pixel. First, one bit determines whether or not the magnetic stylus may alter the pixels orientation. This is necessary so that we don't accidentally overwrite text we wish to preserve. That text could be the documents original text, or it could be previously entered marginalia, in either case we may want to preserve it. Second, in the case of cubical pixels, each pixel needs three more bits to determine what orientation the pixel should take upon receiving a signal from the stylus. In this way we can add notes in a color different from the original text if we so desire. In the upper outside corner of every page we can place a set of six boxes that essentially act as "inkwells" for our magnetic stylus. By "dipping" our pen into one of the inkwells, we can either modulate the signal sent out by the pen to pixels, or simply change the "next state" bits on each writable pixel. By using a special stylus, we can even highlight entire sections of text in a different color, while perfectly preserving the original text.



• Once we have written marginalia or highlighted sections of text, we usually want to preserve these notations. On normal paper, we simply keep every sheet of paper we want to preserve, but as previously discussed, this can quickly become quite cumbersome. Smart paper, allows us to save the pattern of pixel orientations on a given page at any time so that we can store entire page images digitally. Given that each pixel only has six possible states, it is reasonable that each page image wouldn't be any larger in byte size than most current graphical formats. Given even a worst case scenario of image compression, each page image probably requires about one megabyte of storage, which means that we could fit even very large documents into less than a gigabyte. In addition, much of a document would likely be stored as textual information, allowing the smart paper and the reader to make decisions about the fonts used to display that information. Really, the challenge of storing page images is how to connect our smart paper to a suitably large storage device.
  
  First we have to extend our idea of smart paper into the idea of a smart book. By adding a binding no larger than a normal hard-back cover, we gain enough space to contain all the devices necessary to actually make smart paper work. Inside that binding, wireless technology would allow us to store page images in a location remote from the smart paper itself. Wireless network systems currently allow about seven megabits per second transfer speeds, enough that we could load two facing pages in under a second. Given that we can extend the storage capacity of our central page server as necessary, we can effectively expand the storage capacity of our smart book without having to worry about how to lug around giga- or terabytes of storage everywhere we go. Combining our wireless communication system with a RAM of either 16 or 32 megabytes and an intelligent system for predicting what page images might be requested in the future, we can provide rapid response times.



• By providing a special page with a set of command buttons or menus that can be displayed facing any other page of text, we allow the user to interact with the document's digital nature. We can provide search functions in this way. After "pressing" the search button, a reader may either write in a word or phrase, or select a word or phrase from an adjacent page. Again, we don't have to worry about fitting a processor into our smart book because our wireless device sends the request back to a server that parses and returns the requested information. In another instance, a reader may wish to look up a dictionary or thesaurus entry for a word which may then be displayed on an adjacent page or in some blank space of the current page.



• The information displayed on our smart paper is not limited to plain textual material. It is no harder for smart paper to display graphical images than text fonts. Indeed, given the relative simplicity of our pixel system, it wouldn't be all that difficult to add dynamic information. So instead of the series of pictures of an engine we can actually have an animation of an engine with start and stop button. Further, if we wished to browse hyperdocuments, smart books could allow us to easily implement such a system. We already have a two display system of adjacent pages (more if we want to turn the physical pages), and it wouldn't be difficult to make the selection of a link on one page bring up the referenced page on the other display. Again, by utilizing RAM and a system which loads the potentially viewable pages into temporary memory, we can provide rapid response even in browsing situations.

Annotations and Citations