Glaskowsky analyzes the recent XO-2 hardware announcement from the One Laptop Per Child organization.
Peter N. Glaskowsky is a computer architect in Silicon Valley and a technology analyst for the Envisioneering Group. He has designed chip- and board-level products in the defense and computer industries, managed design teams, and served as editor in chief of the industry newsletter "Microprocessor Report." He is a member of the CNET Blog Network and is not an employee of CNET. Disclosure.
The new design uses two touch-screen LCDs flanking a central hinge. This approach allows the unit to be used as a book with facing pages (shown here), as a conventional laptop using a virtual keyboard on the lower display, or as a single system shared by two users.
This is just a conceptual design so far, and the images are computer-generated, but OLPC announced some goals for the redesign-- smaller overall size (about half that of the XO-1), lower weight, 1W power consumption, and a volume price of just $75 in 2010.
These will be difficult goals to achieve, principally because of the doubling of the display area. The XO-1's display is its most expensive component and probably also its heaviest, most fragile, and most power-hungry. As a result, I think the XO-2 design concept is a little too ambitious in a few specific ways.
First, the bezels around the outside of the displays are too narrow. Although modern PC laptops have aesthetically pleasing narrow bezels, this advantage comes at a significant price in the manufacture of the LCDs-- the driver chips have to be mounted on the back of the display glass and connected by flexible circuits. I don't think the XO-2 price target can be achieved without sacrificing the narrow bezel.
Also, LCD touch screens need to be more heavily built than non-touch screens. This is even more true when they're meant to be used by young children, who may use much more force than necessary and may not keep their fingers clean. The XO-2's "keyboard" screen had better be almost bulletproof... and that means more size and weight.
The thin, light, small-outline enclosure also impairs ruggedness. A system intended for use by young children in austere conditions needs to be more heavily built. For example, thin case halves mean thin hinges. Though thin hinges can be strong enough when made with heavy or expensive materials, that solution conflicts with the XO-2's weight and price targets.
The 1W power-consumption target will be especially difficult to hit. The XO-1's LCD consumes about that much power all by itself, and at least twice that with the backlight on. If the OLPC folks mean that the 1W figure is for outdoor e-book reading (CPU idle, backlight off, no network activity), they ought to say that. Giving that figure along with images of brightly backlit displays is misleading at best. (And the color in the simulated images is much better than that available from the XO-1's relatively washed-out display.)
Also, the OLPC people talk about Windows compatibility-- Windows XP, at least-- and it's unlikely there will be any Windows XP-compatible hardware platform capable of achieving a 1W average power consumption figure in 2010... never mind the 1/4W or less that would be available to the processor and chipset with two LCDs running (even without the backlight).
The XO-2 could be based on some cellphone-like chipset, but that would sacrifice Windows XP compatibility. And if there is an XP-capable low-power chipset on the market by 2010, it's likely that it will also be used in more traditional laptops and mobile Internet devices. Similarly, the technology behind the XO-1's LCD is likely to be more widely used in the next few years. In other words, the XO-2 isn't likely to gain any meaningful advantage in this respect over competing platforms.
To me, none of this bodes well for the OLPC initiative in spite of the publicity it has received from these announcements. The organization still seems committed to a strategy of over-promising and under-delivering... not exactly a path to success.
Incidentally, over Memorial Day weekend, I met with several XO-1 users at Baycon, an annual science-fiction convention here in Silicon Valley. This was the largest collection of OLPC users I've ever seen in person. The crowd was a mix of individual users, OLPC developers, and people interested in getting an XO-1 of their own.
OLPC developer Ed Cherlin (shown here on the right in the blue and white-striped shirt) recapped the announcements from last week, described new XO-1 software under development, and answered questions. (I asked him why browser bookmarks disappear after a reboot-- he said this was by design, but that previously visited URLs can be found in the Journal. Sure enough, that works.)
The meeting was also a recap of the shortcomings I've seen in the XO-1. Several of the systems people brought to the meeting had dead batteries because of previous use that day, networking activity while the machines were apparently in standby, or because they hadn't shut down properly. Of the eight or so working machines present, no more than three or four were able to get networked together, probably because of incompatible software versions. Nobody was satisfied with the keyboards. Of course, these were all older users with hands larger than the children for whom these systems were designed, but if OLPC can't get adults interested, there'll never be enough software to meet these kids' needs.
Nevertheless, it was a fun and informative get-together, and it probably persuaded some of the non-owners in the group to get XO-1s of their own when the Give One, Get One program starts up again in August or September of this year.