I thought I should report on where we are with the design of the $50 I-Book Reader. "We", in this instance, being myself, my young business partner Ben Wibaut, and David Thornhill, who teaches engineering at Queens University, Belfast and who is making a model of the I-Reader for display purposes.
Before I list the components we suggest for this machine, with their costs and their estimated power consumption -- an absolutely vital aspect of things for a solar-powered device -- may I remind everyone that its design process is quite different in approach from most information processing gear, including the XO.
We are here asking not how much in the way of function we can supply for a given complexity, but rather the reverse. The only task we have is that of taking a few kilobytes out of memory and displaying them on a screen as a picture in 64 colours or less.
How, then, can we achieve this as simply as possible, and using what in the way of power? The answer here is "with a power demand so small that we can supply it from a solar panel built into the lid of the Reader itself."
We thus, at a stroke, both simplify things enormously and avoid all the costs of outside power provision which bloat the expense sheets for the XO and its competitors.
In the UK, at equinox, a photovoltaic (solar) panel of 8" by 6" can supply a mean of just over 4 watt-hours per average day with a peak output of around a watt. So we should be aiming for an overall requirement for the I-Reader of (say) 0.1 watt minimum, 0.2 watts as a mean, up to a maximum of around 1.2 watts when the processor and other components are running at full transfer speed and the screen is backlit.
This constrains design and performance considerably, but turns out to be achievable. (As always, I invite comments on any aspect of design from anyone who may have them.) Here, then, is a list of all the components needed, with quoted costs where available or estimates where I have been unable to find them.
We start with the MIT (Mary Lou Jepson's) dual mode screen, since we have a quoted price for it at US$16, and known power consumption figures of 0.1W with the backlight off, 1 Watt when backlit
There are several low-power processors we might use, but I have chosen the ARM9 core. It's capable of 300 MIPS when running at 180MHz, though in fact -- again, since all we're doing is taking small pages out of memory and putting them on ths screen -- we probably don't need anything like this speed.
However, it comes complete with 16K each of instruction and data caches, plus 160Kb SRAM and an 8-level interrupt controller, as the A9(!SAM9261 'board', for US$ 10 (quoted, in 1,000 quantity). And it consumes only 0.2 Watts when fully loaded, and a tiny 16 microwatts when in standby -- that is, when a page had been loaded to the screen and its static image is merely being read.
Now that we have solar power in the I-Reader, we only need a small battery. With its charging mechanism, it's mainly used to "smooth out" power delivery from the solar panel itself, plus a small reserve for times when a book is being read after dark. I suggest 2 x standard NiMH AA cells, which will give us 2.4 volts and a capacity of more than 2.4 Watt-hours. Cost in bulk, US$2 at a guess.
With four button-switches, auxiliary power socket, the serial socket (wireless, if possible) for e-book library 'cards', etc: say $11 -- until production volume becomes large enough to reduce this.
8" x 6", sunk into lid as shown here: $4 (quoted price)
Total estimated cost here, therefore, comes to US$43.
Although much of this is merely estimated if not guesswork, I believe that a production run in the hundreds of thousands should easily bring this little machine in at -- as its quoted title states -- the "Fifty-Dollar E-Book Reader."
Note: In point of fact we are currently redesigning the 'lid' so that it is hinged and more closely resembles a book cover, while still containing the solar panel.
In future articles I will try to discuss the actual process by which the Illumination Reader could be prototyped, tested and brought to market, including the unique software which enables it to perform as it does and which, therefore, is crucial to the project as a whole.
I have said -- and I repeat now -- that although I don't believe it will be a good idea to rewrite this piece of software in Linux and thereby make it 'open-source' in formal terms, I am perfectly prepared and indeed intend to make it available free of charge to anyone who wants to create an Illumination book in any language.
In the end (and this applies to the XO as well) it the supply of 'books', of class-work, of information of all kinds which matters more than the actual technology behind that supply. It's the teaching and the teachers, not the blackboard. Professor Negroponte saw that clearly when he set out on this voyage, and many others have observed it too.
We may well argue among ourselves over such matters as whether we need to teach our children -- the children of the world -- computer literacy or just plain ordinary literacy. But, whatever is the case, I believe that this Reader can travel a long way towards the end which Negroponte, and OLPC, identified.
And, therefore, that it will be a very great pity indeed if it is not taken up, supported and developed, for that end and for the benefit of those children who need, first and foremost, words and pictures.