Why Haven't Computers Replaced Books Yet?

Books are one of those things long predicted to become obsolete, to be replaced by an electronic equivalent, but it hasn't happened yet.

We know it hasn't happened yet because everyone is still reading books.

However, computers have partially replaced printed materials, because for many people (including myself), a substantial amount of what they read is presented to them on a computer screen, and not on paper.

The Economics of Technology Replacement

In comparing an old technology A with a new technology B, we might consider the following:

1. Capabilities of A that are better than B
2. Capabilities of A that are worse than B
3. Capabilities of B completely missing from A

In order for A to be replaced by B, we might suppose that A has to be better or as good as B in all respects, i.e.

1. Some capabilities of A that are better than B
2. No capabilities of A that are worse than B
3. No capabilities of B completely missing from A

In practice, replacement happens when the following holds:

1. Some capabilities of A are way better than B
2. Some capabilities of A are worse than B, but they are good enough.
3. Some capabilities of B completely missing from A, but it turns out they don't really matter that much.

For example, there are many things that a horse can do which a car will never do, such as:

• A horse can recognise you as a friend
• A horse can refuel itself from grass growing on the roadside
• A horse can go off the road
• A horse can cross a stream
• A horse can take you up a narrow winding mountain pathway

A horse can do all these things, but none of those advantages were enough to prevent the almost complete replacement of the horse by the motor car as the preferred form of assisted personal transport. I can't claim to be an expert on the history of horse-car substitution, but I imagine that the following are the main reasons why we all use cars instead of horses:

• When not in use, the car takes up a minimum amount of space.
• When not in use, the car can be turned off and mostly forgotten about.
• The car includes a weather-proof enclosure for passengers and goods.
• The car can be locked.
• Once the infrastructure of high-quality roads with hard smooth surfaces has been build, the car provides a much faster and smoother ride from where you are to where you want to be.

Slightly closer to the main topic of this article are examples of replacement of analog technologies by digital equivalents. When the vinyl record was being replaced by the CD, there were claims that when listening to some part of some track on some album, the vinyl record had a sound quality supposedly lacking on the CD version. But even if this was true, the CD had overwhelming advantages of size, convenience and robustness. Similarly, when film cameras where being replaced by digital cameras, it would be claimed that digital cameras can't quite reproduce the high quality of film photos. But the digital camera has the overwhelming advantage that it costs nothing to take a photo and almost nothing to store, reproduce and view photos, and this is enough to render the film camera obsolete as a consumer technology.

Why Books are Better than Computers and Why Computers are Better than Books

Why Books are Better than Computers

• A book can be printed at more DPI than the human eye can resolve (and 600 to 1000 DPI is good enough for most purposes)
• A book can survive contact with small amounts of dirt and water without suffering major damage
• You can drop a book, and it still works
• Books include a "natural DRM", which is the cost of printing. This means that the copyright model works reasonably well, and as a result, the reader can expect to read high quality material written by professional authors who are paid for what they do.
• A book can display at one time an area sufficient to hold 1000 readable words (even if you don't necessarily want to fill all of that space with words in the smallest readable print size).
• Navigation is by direct physical manipulation, i.e. turning pages
• The "user interface" is not overly sensitive, i.e. just touching the book does not cause it to accidentally navigate to the wrong page (or to a different book).
• The book is not attached to anything
• The book requires no power source (although it must be admitted that there is dependence on an existing lighting "infrastructure" if you want to read a book after sunset)
• A book can be held in the hand while the reader is doing any of the following:
  • Sitting in an armchair
  • Sitting in a chair at a table
  • Slouching in a bean bag
  • Sitting up in bed
  • Sitting on a train, plane, bus or car

All of these advantages of books present major technological challenges to the designer of any kind of computerised e-book. But my thesis is that the last item is the only one that really matters, and it is the only thing that stands between the whole book industry and the historical fate of the horse, vinyl record and film camera industries.

Why Computers are Better than Books

• Computers provide access to an unlimited amount of content
• Computers provide access to an unlimited amount of free content
• Computers can provide almost any imaginable interaction model, and in particular almost any imaginable navigational model.

The first two items are very much a function of the Internet and the Web, and like the car example above, show the relevance of infrastructure (i.e., in the case of cars, suitable roads).

The last item is very abstractly described, so it's worth listing some of the specific forms of interactivity that we currently expect to have on our computers:

• Web search engines, and in particular Google, provide immediate access to relevant pages on any topic that can be specified by a set of search terms.
• Wikipedia provides a shared repository of publicly available human knowledge on any conceivable topic.
• Social bookmarking websites provide a steady supply of new, interesting and useful content.
• Any web document can freely link to any other web document.

Computerised navigation can be compared to "book navigation", which has the following features:

• The reader can turn to the next page or the previous page
• The reader can turn pages in bulk, using a variant of the "binary chop" algorithm, to find any page
• A book can contain an index with page numbers for search terms (but only as specified by the book's author)
• The reader can "navigate" to another book by putting the first book down, getting up (from their chair), going to the bookcase, selecting the second book, and getting back into the chair to read it. (For 99.9999% of choices of second book, an intervening step of going to the bookshop to buy it will be required.)

The computerised navigation model seems to have the overwhelming advantage here, with perhaps a caveat in relation to "over-sensitivity" which needs to be considered by anyone designing the ultimate e-book machine.

Book Advantages: Solve Them or Ignore Them

I gave a list above of the advantages that books have over computerised alternatives, which I will summarise here briefly:

• More DPI
• More resistant to dirt and water
• More resistant to dropping
• DRM/Copyright protection
• Large display size
• Hand-operated navigation
• Not-too-sensitive user interface
• Portable
• No batteries
• Hand-held

Some of these advantages are conditional on each other, and in particular on portability. For example:

• A non-portable computer can be placed away from the user (so even if I spill my coffee, it won't damage anything other than the mouse and the keyboard)
• A non-portable computer can have almost unlimited display size
• Keyboards and mice don't suffer too much from "sensitivity" problems, but portable devices require a touchscreen, where sensitivity can be problematic.
• Portable devices have to run on batteries

But if we assume that the computerised must be portable and hand-held, then issues of robustness, size, sensitivity and power supply must be considered by the system designer.

Mandatory Features for the Hand-Held Electronic "Book" (aka Wifi Tablet)

Taking into account the above lists of advantages of computers versus books, the ideal electronic book needs to have the following features:

• Wireless internet (i.e. wifi) and web browsing capability
• Size: somewhere between A5 and A4 (according to user preference, and given existing limitations of screen resolution, with a bias towards the larger size)
• Pixel count: At least 800 by 600
• Battery operated, and rechargeable by swapping between two sets of batteries.
• Touch-screen
• Not including unnecessary features which will compromise weight, size or power consumption. In particular, no keyboard, hard disc, CD drive or speakers. However it may contain features not strictly relevant for straightforward "reading", if they do not cost too much, such as: Bluetooth, USB connectivity, audio output.
• Runs on open-source software, and is user-programmable

We can consider how this design deals with the list of book advantages given above. In some cases the book advantage can be mostly overcome, in other cases we can decide that it doesn't necessarily matter so much.

• More DPI: Typical computer screens run at 100 DPI, which is much less than even the 300 DPI that is considered a minimum acceptable resolution for printed output. However most computer users are quite happy to read content on a 100 DPI screen which is about the same distance as a book might be held (and anti-aliasing helps a little bit), so we can count even 100 DPI as "good enough". New technologies like electronic paper will also help in the future.
• Resistance to dirt, water and dropping Electronic devices are not as robust as simple printed documents, and to make it worse, they cost more, so the user has to be more worried about damaging them. However, the ideal wifi tablet will be somewhat smaller than a typical laptop (e.g. like the screen without the keyboard), which should make it cheaper, hopefully cheap enough to use casually without being too stressed about damaging it.
• DRM/Copyright protection Existing "e-book" vendors have been obsessed with DRM, seeing the e-book less as a replacement for a book from the reader's point of view, and more as a replacement for a bookshop. To the extent that there are e-books which can only read DRM content, and can't read simple HTML. Now that there is more than enough to read on the "free" Internet, DRM becomes a non-issue for most readers.
• Display size: For an A4 screen running at 100DPI, the display size should be large enough. (As I will discuss below, many existing "wifi tablets" put too much value on small size, and the practical experience of reading the web through a small window is too frustrating to be worth the effort.)
• Hand-operated navigation and interface sensitivity If the keyboard is to be left out, then something else has to replace it, and that something is most likely to be a touchscreen. Because A4 size wifi devices are not in widespread use, I don't think that we yet know what is the best model for interactivity and navigation on such a device. Simply transferring the existing keyboard/mouse model to a touchscreen may not give satisfactory results. One frustration in some touchscreen interfaces is that it is too easy to try to do one thing and end up doing something completely different. There is also the problem of reading the device and being stressed out because accidentally touching something causes unintended navigation away from the current page. An important feature of a successful wifi tablet is that it will be user-programmable – this maximises the size of the developer base, and it will be a developer base devoted entirely to solving their own frustrations, which is generally the best kind of developer base to have.
• Portable Having wifi and being battery operated solves the portability problem. But ...
• No batteries Books don't require any power (except, as already mentioned, power for the lights), whereas any electronic device will require some power, and current technology requires a non-trivial size of battery to power an A4 screen and enough electronics to run a web browser and a Wifi connection. So there will have to be batteries. To minimise the inconvenience, the device should have a battery or batteries that can easily be swapped out with a second set (i.e. a second set which you left connected to the charger so that they would be charged when the first set ran out).
• Hand-held This is what I have said is the most important feature that an electronic book replacement needs to have, and it is what determined the design features I have specified, so no more needs to be said.

Text Entry

One additional consideration in the interaction/navigation model is that of text entry. Reading a book does not generally require any text entry at all, however to take advantage of free content on the Internet it is often necessary to enter at least a few words, including:

• User names and passwords
• Search queries

Entering text without a keyboard can be fiddly and frustrating. As with other aspects of interaction, the best solution will come naturally from allowing users to develop their own interaction software according to an open-source model. However I will make the following suggestions:

• Take advantage of the full screen size, and provide a touch screen interface where the keyboard takes up most of the screen.
• Apart from the keyboard, only display the letters currently being typed. Where text is being used to access or discover content, there is little need to see the full screen display at the same time as text is being entered. (In other words, the wifi tablet is not designed to be used as a word-processor.)

What is Out There Now

If the hypothetical wifi tablet that I have described is so good, why isn't someone already making it? Possibly they already are, but as far as I know there is nothing on the market that quite has all the right features.

The first problem to deal with when researching Internet-based "like-a-book" web-surfing devices is what to call them. I have come across the following terms at least:

• wifi tablet
• webpad
• consumer webpad
• couch computer

The set of currently available devices can be divided mostly into three categories, which are:

• Not enough pixels: Nintendo DS & DS-lite, Palm TX, HP Ipaq
• Enough pixels but still too small: Nokia N700 & N800, Vega
• Too expensive: almost anything called a "tablet PC"

The remaining and least populated category is the "almost good enough" category, and there are only two items I know of that are in it:

• iRex Iliad
• Pepper Pad 3

Both of these devices have similar prices (iRex US$699 from "iRex shop", PepperPad 3 currently US$645 from Amazon). The Iliad is based on "electronic paper" with 1024 by 768 at 160 DPI, the Pepperpad 3 has a 7 inch 800 by 480 pixel LCD screen.

As far as software goes, the PepperPad seems to be far superior, as it is a Linux system and you can install a wide range of software (and write your own if you want to).

The Iliad doesn't even come with a web browser (at least there is no mention of such a thing on the product home page), although it is possible to work around this somewhat bizarre omission (for example see here).

Unfortunately I live in New Zealand, and neither of these devices is available in any shops here, so I can only speculate as to how good they are. Most of the reviews for the Pepperpad, such as these and this seem quite enthusiastic.