One Laptop Per Me
Originally published 2006 in Atomic: Maximum Power Computing(heavily rewritten 11 June, 2007) Last modified 03-Dec-2011.
When people talk about The Computer They'd Like To See, they often start banging on about sentient 3D-display neural-interface miracle machines.
Or, at least, mobile eight-core systems with 5Tb solid state drives and 40 inch roll-up screens.
I'd like something a bit more down to earth.
As I write this, the first kids have been been playing with their OLPC XO-1 laptops for some time, though the XO-1 still isn't in mass production yet.
I look forward with interest to getting an XO of my own.
OLPC, in case you've managed not to notice it, stands for One Laptop Per Child. The OLPC project is not really about the hardware, but about getting computing power into the hands of children around the world. But the hardware is nonetheless interesting, because small rugged low power portable computers are good for all sorts of things.
The OLPC XO machines themselves definitely are meant for children, but the only thing about them that's really kid-specific is the rather small keyboard. That's not much of a limitation really, though; if it were Psion-sized then it might be, but it actually seems to be only about 80% the size of a normal keyboard, which is no big deal for most people. And there's a lot more to like in the machine.
A lot of people who buy a laptop computer, you see, don't actually need it.
If a laptop is your only computer and you need to use regular PC or Mac software then yes, that laptop probably does have to be a fully functional PC or Mac in its own right.
But lots of people have laptops as companions to their desktop computer. If that's the case, then the requirements change - as evidenced by all of the people who already use PDAs or fancy mobile phones for e-mail and Web browsing.
The problem with that solution is that phones and PDAs have teeny-weeny screens, their battery life is often pretty ordinary (especially if you're using wireless features), and their keyboards vary from dreadful to nonexistent, unless you buy an add-on keyboard that uses Bluetooth or something.
There is, however, a third class of device, that's been popular in its niche for years now.
Proper keyboard. Built-in non-hinged "mail slot" screen that's big by phone standards but small compared with a laptop. Low power CPU. Non-volatile RAM for storage. Power from AA batteries or inexpensive built-in rechargeables, which give you many hours of run time. Sometimes days on end, or months of occasional use, without a charge or battery change.
Probably the best example of this breed in the retail market at the moment is AlphaSmart's Dana, the "Wireless" version of which has IrDA, USB and WiFi connectivity (in increasing order of interestingness), a 560 by 160 greyscale mail-slot screen (which you can write on with a stylus), and tons of software.
That's because the Dana runs Palm OS 4.1. Any standard Palm app that doesn't need a colour screen therefore ought to work on it. Not all of those apps know about the wide screen (the Dana comes with widescreen versions of the standard Palm apps), so you'll be using a 160 by 160 square in the middle of the mail-slot for many programs. But you can't have everything.
By the weird-operating-system standards of most "PC companions", the Dana is a clear winner.
The only real problem with the Dana Wireless is that it costs four hundred and twenty-nine US dollars. Even the older non-wireless version is still $US350.
You can get a decent used ThinkPad for that.
Now, however, there's the XO-1. And even this first generation of OLPC hardware knocks the Dana into a cocked hat.
The OLPC project's hardware target, famously, was originally a "$US100 laptop". It seems likely to actually be about a $US175 item (in vast government-purchase quantities) at launch, but even that's peanuts compared with other currently available "PC companions".
Most nerds have already heard the basics of the OLPC spec, and seen the round and friendly prototype models with their (now abandoned) crank-handle chargers. The XO-1's got a dual-mode 7.5 inch 800 by 600 colour or 1200 by 900 monochrome screen (colour mode doesn't work in sunlight, but mono does), a 433MHz Geode CPU (giving performance in the same ballpark as a 300MHz Pentium II, about a zillion times faster than the old Z80-driven AA-battery machines that a surprising number of people still use), a built-in camera, wireless networking (802.11 and ad-hoc mesh), and the list goes on. No moving parts, long battery life, quite standard Linux operating system, et cetera.
All this makes the XO-1 a pretty interesting device from the point of view of people like me.
But, as the OLPC Foundation makes perfectly clear, we can all just bugger off if we want to buy ourselves an XO-whatever at retail, at least for the time being.
The above-linked Wiki pages take care of explaining why consumers won't be able to buy an XO-1, or an equivalent device that looks different.
(Retailing the exact same XO-1 model that's being handed out to kids would be an obviously terrible idea, as huge numbers of XO-1s would immediately be taken away from the kids by miscreants who wanted to turn the laptops into US dollars via eBay. This problem's reduced if the retail OLPCs look completely different from the kid models. There's also apparently an anti-theft system built into the operating system, but it doesn't look like a very good idea to me.)
It seems plausible that affluent nerds will be able to buy XO-somethings at a premium price, subsidising their provision to poor kids, at some point in the future. That's been done before, you know - the BayGen Freeplay "clockwork radio" was sold in exactly that way.
The Freeplay radios weren't what you'd call a huge hit, but they also weren't terribly useful devices for us rich Westerners who can easily find, and afford, batteries for our radios. A retail OLPC would be a different kettle of fish.
If the XO-whatever hit the retail market tomorrow, even at three times its wholesale price, I bet the things'd sell in droves. Apparently a fair straight-retail price is more like $US360, though, which'd make a subsidy on top of that price something of a bitter pill.
Whatever ends up happening, though, there appears to be no prospect whatsoever of retail XO-somethings before 2008 - and probably not early 2008, either.
UPDATE: I'm adding this little update in mid-2009, almost three years after I first wrote this column. Things have not gone well for the OLPC Foundation.
We can all thank the OLPC Foundation for helping to start the Age of the Netbook, but the actual XO-1 has turned out to be hideously flawed down to the hardware level. And the Foundation seems to have its own problems, too. It's done outrageous things like apparently not bothering to have a deployment department at all, and doing zero usability testing of the "Sugar" interface, which is therefore somewhat mystifying.
The OLPC organisation expected to have built 150 million XO-1s by the end of 2008. Actually, by mid-2009, they've shipped a few hundred thousand, fewer than 600,000; maybe a third of one per cent of the number they expected to have shipped by now.
Meanwhile, people bought about ten million netbooks in 2008 alone.
The OLPC Foundation is now promising an "OLPC 2.0" that ditches the defective custom hardware. Cynics have observed that the Foundation now seems to have pretty much no technical staff left, so exactly who's supposed to be developing this new laptop is unclear.
And now, back to the original text of this column.
Fortunately, some other companies appear to be picking up the slack. There's been an absolute flurry of PC-companion laptop-ish things announced over the last few weeks (May to June, 2007). Nobody's offering the OLPC's trick screen or human-powered charger, but the rest of the specs - good battery life, no moving parts, wireless, x86 software compatibility - seem attainable.
First out of the gate, at least as far as the announcement goes, was the Palm Foleo. The Foleo's hardware specs are still unknown, though, and it's apparently going to end up costing $US499 after some rebate deal - which is a bit rough for something that fails one of the basic requirements by not being able to run normal x86 software. Palm allege that the Foleo is meant to be paired with a smartphone, not a computer as such; the thing's apparently just meant to be a better input-output device for wireless e-mail users.
So I'm not optimistic about the Foleo being much use as a proper PC companion. But I mention it here anyway because otherwise I'll get letters.
Moving a little further up the vaporware ladder, there's the Via "NanoBook", alleged to cost $US600 and give you a seven inch 800 by 480 screen, four hour battery life, and the raw animal power of a Via C7-M CPU.
That sounds... well, not good exactly, but better than the Foleo. Except the NanoBook's only been shown as prototypes so far, with retail versions lost in the mist of the future. Perhaps as soon as August, but I'll believe that when I see it.
At this juncture, it might be worth mentioning that the entry level price for a brand new proper PC laptop from Dell is $US549. Well, it is if you buy when Dell have a cut-price deal going, which is of course the only time when you should buy anything from Dell, the dodgy Persian carpet sellers of the PC world.
A $US549 Dell laptop is not a very exciting product - it'll have half as much RAM as it needs, for a start - but it's still worth remembering. Heck, you can get a Lenovo 3000-series laptop for $US649, or a (similarly Lenovo-built) new ThinkPad for barely more. The particular charms of a PC companion start to fade when the darn things cost as much as these "proper" laptops.
Which brings me to the big news in the PC-companion market: Asus' announcement of the piercingly-named "Eee PC 701", which is also alleged to be coming in August.
The Eee is the first example so far of Intel's Classmate PC platform, the specs for which aren't amazing. Seven inch screen, 512Mb RAM, 4Gb Flash-RAM storage.
But that spec level is supposed to only cost $US199.
For something with an x86 CPU (apparently it'll come with some flavour of Linux, though it can also run WinXP), that's a big deal.
The Eee PC is actually - according to the marketing spiel, anyway - aimed at the same sorts of markets as the OLPC systems, which explains the low price (and considerable hatred from the OLPC gang). But it'll apparently be sold through normal retail channels. So you won't have to wait for a special version to hit retail, or steal one from a doe-eyed Third World child, to get hold of one.
Over the years, quite a lot of other PC-companion sorts of devices have cropped up. Quite a few writers still peck away at Cambridge Z88s. An old Palm organiser with an add-on keyboard will get the job done for many users, some of whom are not crazy.
I myself have an Amstrad NC100 (one of the old guard of Z80 machines) sitting on a shelf. An NC100 is still A-OK for many purposes, especially if you just want something to write with. Provided it, you know, works, which mine doesn't.
And, as my blog readers already know, I've recently reconditioned an Apple eMate 300, which is now waiting for me to solder up a serial cable so I can start bootstrapping software onto it to let it connect to a PC in some more elegant way.
If you're not quite that dedicated, then weird old things with mail-slot screens and processors that chug along below 1 MIPS aren't likely to interest you, no matter how cheap they are. You'll want something like the Eee PC, with a standard Web browser, WiFi and direct compatibility with modern file formats.
And heck, for $US199, I want one too, tediously unquirky though the Eee PC looks likely to be.
But I also know that ten years from now, one way or another, I'll have an OLPC product sitting on the shelf where that eMate is now. And it'll only be a few solder joints away from working.
