Sunday, January 06, 2008

OLPC First Impressions

A few weeks ago I got my first chance to try the OLPC XO-1 (One Laptop per Child Children's Machine or $100 computer), sub-notebook PC. We had a lineup of two different pe-release models of the OLPC and an ASUS Eee PC for comparison. Overall the OLPC looks a solid device, but encumbered by an experimental interface and a paternalistic view of education for developing nations.

My first impression was of how bright and robust the green plastic case looks, like a toy for toddlers. The bright green rubber keyboard is well laid out, but the keys are too small for adult fingers. The very wide touch-pad under the keyboard makes it difficult to find somewhere to rest your palms.

The OLPC uses the new Sugar child friendly graphical user interface, making it unusable to those familiar with the desktop metaphor of Microsoft Windows, Apple Mac or various Linuxes. Sugar may well be a better user interface for children, but will have to be learned by anyone familiar with current computers. The OLPCs used their WiFi mesh network to recognize each other automatically, representing the other machine on the screen graphically, but due to lack of familiarity with the interface I was unable to use the connection.

The screen on the OLPC looks very much larger than the 7 inch wide LCD on the Eee PC. The transflective OLPC screen did not look as bright or clear as the Eee PC indoors. Outdoors on an overcast day, the OLPC screen switched to monochrome mode and was readable, but still not a lot more so than the conventional back lit Eee screen.

Overall the OLPC hardware looks solid and the software usable. However, the unit may be breaking too much new ground to be successful as a product and is hampered by being associated with a flawed model of development.

The users of the computer will have to learn a new user interface different from that predominating. It may well be a better interface, but it took decades for the desktop metaphor to be refined and become established. Xerox's interface was adapted for the Apple Lisa computer and then simplified for the Apple Mac before it became popular (and then adapted, or some would say degraded, for Microsoft Windows). It may take ten years and two more teams of designers before the OLPC interface is ready for widespread use.

If the OLPC was purely a university research project, or a product funded by private venture capital, then it would be worth taking a risk to develop a revolutionary new computer. However, the OLPC is intended to be used for education in developing nations. Therefore the resources spent on the OLPC have to be diverted from other education projects for developing nations. In effect the OLPC will take several hundred dollars away from each child in a developing country and using that money to conduct an educational experiment on those children. Developing nations, and parents, may have other ideas of educational priorities for their children and of how computers could be used to assist development.

It may be useful to decouple the various parts of the OLPC and develop them separately. New hardware can use new or old user interfaces, educational content can be developed to work on new or old hardware. Models for funding education for developing nations can be developed separately of a particular educational products. The customers can then assess how successful each of the products is and which they wish to use.

1 comment:

Tom Worthington said...

My first impressions of the OLPC XO-1 were not all that positive. For a more favorable review and details of the hardware, including photos of the components, see bunnie’s blog. They show for example, how the LED backlight for the LCD screen can be removed.

This got me wondering: given the back-light uses much of the power, could the back of the case be made translucent so it would work with ambient light instead of power? This would not be feasible with the OLPC XO1 as the main circuit board is located behind the display, but would be possible with a conventional laptop or desktop LCD screen. Some sort of reflector might be needed to collect enough light, but this might be a way to make full color LCD screens which would work in full sunlight.

This is not a new idea, I recall having a diecast model car with "working headlights". Inside the body of the car was a piece of transparent plastic which directed and concentrated light from the rear window to the front "lights". The result was quite bright.

Something similar could be done behind an LCD display, using reflectors. The result might end up being about the size and shape as an old fashioned CRT screen, but could be made to fold up for transport.