XO-2? No thanks


I was pretty disappointed to hear that OLPC was going to announce new hardware a few weeks ago. The XO was finally shipping and OLPC had a plan to start the long, hard work of getting the software up to snuff. But instead of using their great, deliverable hardware as an advantage OLPC has decided to throw it all away for a few minutes in the spotlight.

Sad XO-1
The XO-1 Gets No Love

The XO hardware is a lot like a game console -- it's a fixed hardware platform where every one is exactly the same and every unit that rolls off the assembly line makes the next one a little bit cheaper. Customers who buy one the first day it's available can run the same programs as those who wait a few years until they're a little cheaper. The best part is that new software just gets better and better without the hardware ever changing.

See, as developers work with a fixed platform they really cozy up to the hardware, learning the tips and tricks that make their software maximize the hardwares performance. You can see it in just about any console game with a roman numeral in the title or any annual sports game franchise; every year the programs look and play better and
better, even though it's the exact same chips running things as the year before. It's all in how the software is better crafted for the hardware it runs on, and that takes time.

This is exactly what I was hoping to see with the XO. The XO-1 hardware is great. But the software still leaves a lot to be desired. OLPC made a roadmap specifying a couple of minor software updates and then a new generation next year. This was encouraging, OLPC looked like it was ready to do the hard work necessary to finish the job and
make the XO a great educational tool, not just a flashy tech demo.

But now here comes this XO-2 thing. Why would anyone buy an XO-1 when there's a new one on the way? If the hardware's better, countries considering the XO will just wait until the new one is ready, seeing no reason to saddle themselves with a million old laptops. Sneaky countries might even just order the next generation to grab some good press on the cheap, knowing they can quietly change their mind by the time it's released. And countries that need others to buy first to bring the price down will have to wait that much longer for their opportunity to implement laptop initiatives.

I expect that XO-1 software development will wind up looking a lot like the homebrew scene for Dreamcast game development. Or Amiga or Atari 2600 or whatever your dead platform of choice is. The XO-2 announcement makes the XO-1 look clunky and uncool to write software for, especially if there won't be any customers. And there's no
financial incentive for anyone but OLPC employees to write software. Even OLPC workers won't have much reason to do anything great -- instead of doing the hard work to get a fast web browser they'll just wait for the next generation of hardware. We've seen that happen XO development before, when development for the slower B2 test machines was dropped once the B4 test units were delivered. That time only volunteer software developers were left behind, but this time it's OLPC customers.

So that's why I'm disappointed in OLPC and their plans for the XO-2. They could have just mildly improved their great, shipping technology and built on their work in the field. Instead they've thrown it all away for another demo and press release. G1G1 participants, get comfortable with those activities in the shipped XO-1 build 650 because they aren't going to get much better.

Jonathan Blocksom is a software developer living in the Washington, D.C. metro area. He is active in the OLPC DC Learning Club.

Related Entries


>They could have just mildly improved their great, shipping technology and built on their work in the field.
They are.. making it cheaper to produce and field repair, right? There was a newspost here a couple weeks ago or so..

I had an Apple ][ (still do). Had a lot of fun with the tips and tricks on it, tweaking software to maximize the hardware. But I'm glad Apple moved on to the Apple IIe, Lisa, Macintosh, iMac, etc., etc., etc.

I will put my G1G1 XO-1 laptop next to my Apple ][ and move on to better things with the XO-2 when (not if) it comes to that. And I will be glad for that too.

OLPC has little to no business sense. By announcing the XO-2 so far ahead of its release, they're just killing potential sales of XO-1's. This is precisely the reason why companies like Apple are extremely quiet about products updates until they are ready to launch them.

The author is absolutely right.

And why should anyone believe that the software for the XO-2 will be any better if OLPC can deliver only crap for XO-1. We are still waiting for Update-1 and it is nowhere in sight !!!

>We are still waiting for Update-1 and it is nowhere in sight !!!

I've been running the -faster and -joyride builds for a while now, and they're getting decent with respect to features, stability, MUCH faster activity load times, and all-round user experience. It's nice to close the laptop and see it go to sleep, and open it and have it fully running before you've set the screen to the right angle. The slightly revamped sugar interface gets more and more stylish with each build.

People seem so impatient (I am too, so I run unstable builds :P).

You guys are idiots. Didn't you get the memo? Only countries that have bought the XO-1 can buy the XO-2. Make sense now?

> Why would anyone buy an XO-1 when there's a new one on the way?

Indeed: http://en.wikipedia.org/wiki/Osborne_effect

Nihil novi sub sole.

re QuantumG: on the contrary. The countries already having some XO-1 are the most probable customers -- by the time XO-2 is available they will have essentially run pilot programs and amassed experience/best practices etc. they can further use.

Also how is price lowering due to economy of scale supposed to work when part of potential sales is cannibalised by XO-2?

@QuantumG: That's even more wrongheaded if true -- you're artificially restricting your already highly artificially constrained market (which has been a problem with the XO-1 sales), plus giving them no reason to work on the XO-1 implementation since so much could change with the XO-2.

Cannabilism of existing technology is always a problem but the XO-2 is two years away at best so it shouldn't affect anyone's deployment plans. It's ridiculous to think that OLPC wouldn't continue to improve its hardware. The XO-2 is a bold step forward, at least in concept. I have doubts about their ability to bring it to market at anything like a $75 price point but that's a topic for another discussion.


Update-1 has been out for some time. It's called Release Candidate 703. The only reason it's not official yet is because all the required testing is not yet complete. Go to the OLPC Wiki. There's a green box on the upper right of the home page that contains links to the latest stable build and the current Release Candidate. Click on Release Candidate (not 703) and you're sent to Update-1 release notes. If your interested in the future, click on the roadmap link there.

Yes, there is the Osborne effect, but olpc said it was releasing the plan for the XO-2 so early partly because it was hopping it would stimulate oem's to adopt many of the ideas. Remember, this is not a for-profit business.

There are a number of reasons for the X0-2. These include cheaper production costs and lowering the power requirements so kids could use the yo-yo.

As far as software development goes, my guess is they will keep the basic circuitry similar enough that it will not be too hard to upgrade applications. The innovation seems to mainly be elsewhere in the hardware.

simple solution, platform-independent software. Good now, later, good for Eee, good for legacy software, good for recycled computers.

> The XO-2 announcement makes the XO-1 look clunky and uncool
> to write software for, especially if there won't be any customers.

Wouldn't you know! As soon as I comment about Release Candidate 703 being the prospective Update-1 build, they make it official today! See changes to green box on OLPC Wiki site.

erj, take a closer look. 703 is now update.1 but still contains the bug (#6532)that wipes out SD cards. They couldn't resolve the problem so they just released it with the bug. This is NOT good.

Why don't you help the software development rather than just talking?

One more thing. this is not for you, media, or rich people but design for poorest children on earth. SO, you have to sacrifice computing power to lower the price and cost.