OLPC XO-3 To Debut At CES, Starting Under $100 (But Not For You)

Computerworld is one of many publications heralding the expected arrival next week of the long-awaited OLPC tablet, and making much of one very cool feature: the price. The initial XO laptops from OLPC never quite made it to the hoped-for under-$100 level. But at least with an ordinary LCD screen, says project founder Nicholas Negroponte, the new XO-3 actually has. (An optional daylight-readable Pixel Qi screen bumps the price up, but it's not clear quite how much.) Both OLPC and Pixel Qi will be at next week's CES; hopefully I'll get a chance to provide some first-hand details, and ask whether there will be another round of the Buy One Give One program, so users outside the reach of big government buying programs can both further the project and play with the product; so far, the word is that these will only be available for large government buyers. (TechCruch has better pictures of the new device.)

They're too late and way overpriced for the market

[tomhudson]

$52, but throw in a $20 government subsidy and people are getting them for $35. each. How a Montreal company won the race to build the world's cheapest tablet [theglobeandmail.com] - it runs linux and android - the cost - $52 each. Here's just a small part of the story.

Published Thursday, Dec. 29, 2011 6:40PM EST

In the morning, Suneet and the remaining three bidders return to the same room. At the front, a 12-person committee shows off the submitted tenders, time-stamped and sealed with wax, before reading off each companyâ(TM)s bidâ"including the lowest estimate of what it would cost to make the Indian governmentâ(TM)s dream: the cheapest tablet in the world.

When the presentation is finished, Datawindâ(TM)s price tagâ"$52â"is the lowest. The next cheapest bid is for $64. âoeI went white,â Suneet says now. âoeI thought, âWeâ(TM)ve missed something.â(TM)â

Feeling nauseous, he staggers out into the antechamber, where rival bidders lob wisecracks in his direction. âoeAt that price, weâ(TM)ll buy some,â one businessman says, laughing. Frantic, Suneet calls Montreal, where it is nearly 3 a.m., knowing heâ(TM)ll wake up Raja. But his elder brother, who at times forgets how many patents he has to his name (more than 50) but never forgets product specs, reassures him that the final price accounts for every single component in the device. Thatâ(TM)s when it sinks in: Theyâ(TM)ve nailed this.

So far, Datawind has manufactured about 10,000 of its ultracheap devices, and has subcontracted more factories in India to gradually churn out a volume of tablets that still seems unbelievable to the founders. The Indian state plans to subsidize the tablets down to between $20 and $35 (U.S.), to be sold to college and university students, and wants to roll the devices out to around 12 million users over the next 12 months. After that, the goal is to place one of these tablets in the hands of each of the countryâ(TM)s 80 to 100 million high school students. The process, despite the hype, is still in a nascent stage, unfolding slowly.

But things got stranger. Shortly after the announcement, Suneet was invited to meet with Thailandâ(TM)s Minister for Information Communications Technology (who was so interested in purchasing 10 million tablets that he attended their meeting even as flood waters descended on Bangkok). Calls arrived from Turkey (which wants 15 million tablets), Sri Lanka, Trinidad and Tobago, Panama and Egypt. At one point, the Swedish embassy in Canada called: Would Suneet possibly have time to meet the Swedish foreign minister, Carl Bildt? And would it be possible to send out a press release to announce that the meeting was happening?

Another story: from pcworld [pcworld.com]

And for an extra $10, you get a much better cpu, a better touch screen, more battery life, etc.

So, forget Canonicals' secret plans to unveil a cheap tablet running linux next week - these run both linux and android, and they're already being sold.

The first OLPC overpromised and underdelivered

[dpbsmith]

I was one of the original G1G1 participants, and I'm sorry to say that the gap between what was promised and what was delivered would never have been forgiven in any commercial enterprise. The "20 hour" battery life turned out to be 3-4 hours, and despite much talk about improvements to the power management software, nothing ever came of it.

The biggest disappointment for me was that the much-heralded "show source" button, didn't. I never quite worked out the tortuous explanations/excuses, but one of the original premises was that all of the machine's source would be available for inspection and modification--to kids, if sufficiently bright. In reality, all the enthusiastic video demonstrations of the "show source" feature were just showing ordinary browser HTML source, and as nearly as I could tell, the "show source" button never did anything more than that.

"Sugar," which I'd hoped would educate me in a brand new model for computer interaction, was, at the time, a bad joke with poor usability. The only way to locate journal entries was by remember to enter text tags for each one when complete, and doing text searches on the tags. It was explained that "fortunately kids like to describe everything they're doing." All usability objections were answered with the retort that I was not part of the machine's intended user base--true enough, and I have never verified for myself whether eight-year-old kids using the OLPC laptop really do type in text tags to enable them to locate their documents.

The one practical use I meant to put it to, as an eBook reader for PDF documents, didn't work because the PDF reader program was buggy, crashprone, and--even when it didn't crash--didn't save your place in the document (and didn't have any bookmarking mechanism). If you stopped reading at page 56, when you reopened the document, you'd be at page 1 and would have to remember what page you were on and scroll to it.

Hopefully all of these problems have long since been dealt with, but it left me with a bad taste in my mouth.

Re:Mass production

[grcumb]

In manufacturing it's all about volume. If you make 10 times as many the price per unit drops by half or so. Make it and sell it everywhere.

You're ignoring the costs of marketing[*], supply chain management, vendor relations, legal compliance. technical certifications, tax and tariff issues, etc. etc.

If you're an existing computer seller (e.g. Dell, Lenovo), you've already got a significant investment in these areas, but if you're a small organisation whose target is the developing world, bootstrapping a global distribution network might seem like a distraction.

Of course, there are a number of ways to work around this, like forming a strategic partnership with a large distributor. But if history is any example, the large vendors are anything but enamoured with OLPC. Nonetheless, there are ways to achieve what you describe. I just don't think they're as trivial as you make them out to be.

-------------
[*] I don't mean cheesy advertising shills, I mean marketing in the sense of determining how the whole supply chain is going to be managed, figuring out who to talk to, what volumes to anticipate, etc.

If you really want to get your hands on one

[object404]

Hey everyone.

Although units are very hard to get a hold of, if you're really sincere and interested about developing, OLPC will ship and lend you units free of charge with the promise that you will pass them on to the next developer when you're done with your project.

msobkow, all you need to do is to make a good project proposal and apply for the contributors program:

http://wiki.laptop.org/go/Contributors_program [laptop.org]

They really do send these out. I applied and OLPC sent over some units all the way to the Philippines

You guys can check what's happening with the different OLPC mailing lists here:
http://lists.laptop.org/listinfo/ [laptop.org]

And the developer mailing list which is the most active:
http://lists.laptop.org/listinfo/devel [laptop.org]

I've also been able to do some hands on testing stuff on a prototype XO-1.75 which is the Marvell Armada-driven ARM version meant to succeed the XO-1.5 (as well as being the basis for the XO-3). It's been a really interesting experience with the prolonged battery life, but not without its quirks as a "real mainstream linux" OS running on an ARM machine (it's running Fedora ARM, dual bootable to the Sugar UI paradigm or Gnome). If anyone wants to contribute to Fedora-ARM development, this would also be an excellent avenue.

Try to check if there any local groups near your place and check em out. The local group near where I'm at right now (NZ) was kind enough to lend me one of these rare prototypes (and will be returning it soon).

Cheers!

-Naz

for education, still open source

[spage]

OLPC's customers are educational organizations that can implement "one laptop per child".

A lot of the OLPC software effort is easing the hard work of a deployment [laptop.org]: managing reflashing hundreds of machines at once with a new distribution, restoring to a stable image, device backup, school servers, service & repair, etc. That's more involved than "selling low-cost computers" and it's different from "the democratization of computers". Android and ChromeOS have some similar facilities and someone could base large educational rollouts on them, but there's little money in it, so it seems if a non-profit is still the way to go.

You're confused (or writing poorly about fish). OLPC never "jettisoned" Sugar. The OLPC software distribution now offers a choice between the Sugar UI and a Gnome desktop, and supports running a version of Windows XP from SD card; OLPC provided these choices in response to those education customers. Of the 2.5M XO laptops out there, no large deployment is running Microsoft Windows. In many Sugar activities, pressing View Source (Fn + Space) opens up the Python source code (it's pretty cool!), and the source code from the firmware up is readily available.