Sugar on a Stick v1 Strawberry is Out

Although I've been sceptical of the OLPC project, not least because of its ridiculous decision to offer a Windows XP version - putting them in thrall to a US monopolist is really good way to "help" the developing world, people - I'm a big fan of the GNU/Linux-based Sugar Learning platform. I'm also a big fan of using USB sticks as a way of providing complete software solutions based on GNU/Linux. So it should come as no surprise that I think this is fab:

Sugar Labs, nonprofit provider of the Sugar Learning Platform to over one-million children worldwide, announces the immediate availability of Sugar on a Stick v1 Strawberry. Available free for download at www.sugarlabs.org, Sugar on a Stick can be loaded onto an ordinary 1GB USB flash drive and used to reboot any PC or netbook directly into the award-winning Sugar environment. It runs on recent Macs with a helper CD and in Windows using virtualization. Sugar on a Stick is designed to work with a School Server that can provide content distribution, homework collection, backup services, Moodle integration, and filtered access to the Internet. Today’s Strawberry release is meant for classroom testing; feedback will be incorporated into the next version, available towards the end of 2009.

...

Learning Activities are at the heart of Sugar. Sugar on a Stick includes 40 Activities to interest young learners such as Read, Write, Paint, and Etoys. Hundreds more Activities are available free for download at the Sugar Activity Library. Most “Sugarized” Activities have student collaboration built-in; students and teachers work, play, and learn on the same Activities together. The Sugar Learning Platform is open, so by leveraging the work of other open source projects, existing software for children can be integrated; for example, the acclaimed GCompris suite of 100 Activities developed over the past five years by Bruno Coudoin was recently added to Sugar, including Activities such as Chess, Geography, and Sudoku. Teachers and parents interested in Sugar’s Activities and its modern interface for children can watch short videos on the recently opened Sugar Labs Dailymotion channel.

Note that this is a great way to (a) use old PCs (b) provide educational materials in a single package for free (c) to avoid security issues associated with Windows. What's not to like?

Follow me @glynmoody on Twitter or identi.ca.

Nobody Buying Windows XO Laptops?

One of the most vexed questions at the moment is the state of the netbook market. What are the returns for GNU/Linux-based systems? And is it true that Windows XP is all-conquering here? Unfortunately, despite all this sound and fury, there are few data points to provide much guidance. So some news from the the OLPC community is particularly interesting.

As you may recall, one of Microsoft's bigger "victories" was getting OLPC to offer Windows XP versions alongside the original Sugar-based system, running on GNU/Linux:

It was almost exactly one year ago that Nicholas Negroponte announced an agreement between OLPC and Microsoft to bring Windows XP to the XO-1 to great turmoil. I vividly remember the late-night flood of e-mails and IRC chats where everyone was trying to figure out just what that announcement really meant.

Someone from the One Laptop Per Child News site has had the excellent idea of following things up:

I've been wondering about what ever happened to these Windows XP-based OLPC trials. I haven't really heard anything about them in quite some time. Now more recently I've asked around and found there is a good reason why I haven't seen anything: countries are choosing Sugar over Windows XP for their XO deployments.

Apparently the conversations are going pretty much as many of us had expected: Initially country representatives inquire if Windows XP runs on the XO laptop. That doesn't really come as a surprise - for many people Windows is the definition of a computer. However, upon further investigation every country decided to stick to Sugar.

So, is this evidence that XP is a damp squib in the OLPC world? And what implications does it have for netbooks?

OLPC: Out; OLPH: In

As regular readers of this blog may have noticed, I've given up on One Laptop Per Child. Happily, I've now come across something to fill the meme-sized hole that leaves:

Gdium.Com is launching the One Laptop Per Hacker program.

Here is your opportunity to contribute to the Gdium revolution.

The Gdium Team is opening a site and a program centralizing all the developer centric resources for the Gdium.

The OLPH program is supporting developers, contributors, creative artists, and other innovators who wish to:

* Optimize, improve the OS, Human Interface and/or Application stack of an “education centric” netbook.
* Experiment with the look and feel
* Provide and disseminate their new application stacks
* Redesign the artwork
* modify the hardware or integrate some nifty gadgets.
* Experiment with the Gdium to support new vertical markets.

For a limited set of selected contributors the Gdium Team provides a set of materials and services, enabling them to get an early start on this machine.

Gdium, in case you were wondering (as I was), are creating the groovy Gdium Liberty, which rather idiosyncratically uses Mandriva (remember that?).

How the OLPC's Rose Got its Canker

This blog post explains in painful detail how OLPC was "turned" by Microsoft - and hence why I have personally given up on the project:

As part of a small personal project, I've been reading through the court exhibits presented in Comes V Microsoft. One of those exhibits is a chain of internal Microsoft emails discussing how to get Windows XP on the OLPC.

...

Finally, in case you think I've failed to mention it: there is never any talk of "the best technology" or "educating or empowering children" or "customers/governments want Windows" or any such merit-based discussion. Outside of a brief mention of Academic Software offerings - literally the very last thing in the recap and suggested by the OLPC faction - the entire discussion revolves around what benefits Microsoft, what might hurt Google, and exploiting inside information they have on the OLPC project and OLPC people.

Read it and weep.

Give One, Get One: I Still Don't Get It

There can be few open source projects that offered so much promise, and yet which have so signally failed to deliver, as One Laptop Per Child. As I noted below, open source software seems made for education, and the idea of combining that with hardware specifically designed for children in developing countries, with all that implies in terms of ruggedness, power availability and access to infrastructure, seemed just inspired....

On Open Enterprise blog.

Revenge is (Sugar) Sweet

On Open Enterprise blog.

OLPC is Dead...

...and Matthew Aslett is dead-on:

“One can be an open-source advocate without being an open-source fundamentalist,” Negroponte told the AP, while lamenting that the focus on open source software had caused technical problems, such as limiting support for Flash. “Negroponte said he was mainly concerned with putting as many laptops as possible in children’s hands,” added the AP.

The focus on laptop sales is laudable, but it is debatable whether it justifies abandoning open source software. This is a matter not of fundamentalism, but of principles.

Sad, but the prospect of Sugar running on other low-cost GNU/Linux laptops almost makes up for it.

Update: Even more on Negroponte's insane embrace of Windows XP, and his apparent lack of understanding as far as open source is concerned, here.

Open Enterprise Interview: Mary Lou Jepsen

On Open Enterprise blog.

Saving Limbu from Linguistic Limbo

There's a fair amount of acrimony flying around the OLPC XO machine at the moment, which is a pity. Because the real story is stuff like this old but still important post that I came across recently:

The development team at OLPC Nepal have been working hard on developing various learning activities for children using the XO. A significant area in which they have been making progress has been in creating activities to help children learn their local dialect.

The first dialect to be setup for use on the XO is Limbu. This is a Tibeto-Burman language spoken by more than 300,000 people in eastern Nepal as well as parts of Myanmar, Bhutan and India.

This is a really exciting development and is a positive counter to concerns that the OLPC project will only serve to homogenise indigenous cultures. In fact, the project may aid the long term preservation and viability of minority dialects and culture which are no longer part of the curriculum in the traditional school teaching models.

I always was a sucker for those Tibeto-Burman languages....

OLPC's Founding CTO Mary Lou Jepsen

OLPC's XO has been much in the headlines recently. If you want to find out it all from someone who knows, there's a good interview on Groklaw with the ex-CTO, Mary Lou Jepsen. And no punches pulled, either:

Q: The world is now very aware of the spoiler role that apparently Intel tried to play. Can you though talk to us about the differences technically between the Classmate and the XO?

Mary Lou Jepsen: Where to start: Classmate is more expensive, consumes 10 times the power, has 1/3 the wifi range, and can't be used outside. Also, the Classmate doesn't use neighboring laptops to extend the reach of the internet via hopping (mesh-networking) like the XO does. So not only is the XO cheaper than the Classmate, the XO requires less infrastructre expenditure for electricity and for internet access. In Peru we can run off of solar during the day and handcrank at night for an additional $25 or so per student – this is one-time expense – the solar panel and the crank will last 10 or perhaps 20 years. Just try running electricity cables up and down the Peruvian Andes for that cost while making sure it's environmentally clean energy. The Classmate isn't as durable as the XO, and its screen is about 30% smaller, the batteries are the type that can explode and only last 1-2 years and can't be removed by the user and harm the environment. The batteries are expensive to replace: $30-40 per replacement. The XO batteries last for 5 years and cost less than $10 to replace. Finally, the XO is the greenest laptop ever made, the Classmate isn't – this matters a great deal when one proposes to put millions of them in the developing world.

OLPC XO-1 Exposed

The OLPC XO-1 project has been something of a roller-coaster ride. Widely praised when it launched, and then progressively pooh-poohed (including by me) as the wrong solution, more recently praised once more for the execution of the underlying idea, and finally put in doubt following Intel's rather suspect ship-jumping.

For what it's worth, in the light of the reality, rather the idea, I've changed my view: the OLPC XO-1 really seems to serve its intended audience well, and to be a well-desinged bit of kit. Confirming that is this splendid deconstruction by arch-hacker Andrew "bunnie" Huang , founder of the similarly-spirited Chumby project:

If I were to make one general comment about the OLPC XO-1, it’s that its mechanical design is brilliant. It’s a fairly clean-sheet redesign of traditional notebook PC mechanics around the goal of survivability, serviceability, and robustness (then again, I’ve never taken apart any of the ruggedized notebooks out there). When closed up for “travel”, all the ports are covered, and the cooling system is extremely simple so it should survive in dusty and dirty environments. Significantly, the port coverings aren’t done with rubberized end caps that you can lose or forget to put on–they are done using the wifi antennae, and the basic design causes the user to swivel them back to cover the ports when they are packing up the laptop to go. That’s thoughtful design.

If you're interested in what makes innovative hardware tick, don't miss this fascinating exploration. (Via Boing Boing.)

Fearful Symmetry

I've noted before that Microsoft is in difficulty over the ultra-mobile machines like the Asus EEE PC; now it seems that the other half of the Wintel duopoly is also in trouble because of the new triple-P (price, performance and power) demands these systems make:

Two days before Intel CEO Paul Otellini would unveil the Classmate 2 or the Intel-powered XO at the CES, Intel announced that they are quitting the OLPC board.

Intel claims that they are quitting because of Nicholas Negroponte wanting them to stop the promotion of the Classmate/Eee to education in third world countries, but I think that the real reason is that Intel does not have a good enough processor for the OLPC project to use as an alternative to the AMD Geode LX-700.

Intel has not been able to develop a processor to match the price, power consumption and performance requirements of the OLPC project.

Bye-bye Wintel, hello Linmd?

Whatever the Question, the Answer's GNU/Linux

It's interesting that whenever people try to come up with low-cost machines for developing countries, the answer is GNU/Linux. The OLPC/XO is the best-known example, but here's another one:

Rather than one inexpensive laptop per child, the answer being presented is a somewhat more powerful computer, with zero maintenance or moving parts, which can be shared by a number of children running free and open source software.

Of course, it's pretty obvious why: the cost of software is zero, which means there's no money there wasted on fat cats in Western countries. But there's another interesting angle:

Perhaps the oddest technical feature of the tablet-style PC is the fact that it runs an ARM-based RISC CPU, the Freescale i.mx31. The key reason for choosing this RISC CPU over a conventional x86 Intel or AMD processor was battery life. Morgan explained that this CPU's power envelope of just 3.5 watts made an 8-hour use possible. The other key reason was that this particular chip had strong video and graphics capabilities, which would be needed to show videos and animations in a classroom environment.

The downside is that it does not run the vast amount of x86 software out there.

The operating system is a cut down version of Debian Linux, recompiled for the ARM architecture, complete with most of the office and communications software expected in a GNU/Linux system.

Oh look: there's that wide platform support again: is this turning into one of open source's best-kept secrets? (Via Linux Today).

Oh, Well Done, Microsoft

Look: little Johnny Microsoft is doing *ever* so well in his plucky attempt to catch up with that clever GNU/Linux chap:

Microsoft Corp has made progress in getting its Windows software to work on a low-cost laptop computer for poor children that currently runs on rival Linux software, an executive said on Thursday.

The world's largest software company is now working to adapt a basic version of Windows XP so it is compatible with the nonprofit One Laptop per Child Foundation's small green-and-white XO laptop.

"We're spending a nontrivial amount of money on it," Microsoft Corporate Vice President Will Poole said in an interview on Thursday."

But be warned:

"We remain hopeful with our progress to date, we still have significant work ahead to finalize our analysis and testing processes," he said. "At the end of the day, there's no guarantees."

So, just remember that: when you're dealing with Windows XP, there are no guarantees. Unlike with GNU/Linux, of course, since it runs rather nicely on the XO already. Now, which would *you* rather have?

Good News, Ulaanbaatar

For my one reader in Mongolia, good news:

President Nambaryn Enkhbayar of Mongolia announced today his commitment to provide every child in his nation with a connected laptop by the end of 2010.

As a first step toward making this a reality Professor Nicholas Negroponte, founding chairman of the non-profit association, One Laptop per Child (OLPC), and Mr. Nyamaa Enkhbold, Minister for Foreign Affairs of Mongolia, agreed to launch the OLPC initiative in Mongolia as early as January 2008 and signed a Memorandum of Understanding (MOU) thereof in the presence of President Enkhbayar.

In January, Mongolia and OLPC plan to launch a pilot project providing 20,000 units of the acclaimed XO laptops, to children ages 6 to 12 in the most remote parts of the country, as well as in the capital Ulaanbaatar.

I'll get me yurt....

The Coming Java Tsunami

I think this is just the first of many such decisions, all born of Sun's enlightened choice of the GNU GPL for Java:

Python was originally the language of choice for OLPC [Open Laptop Per Child] but with the announcement of the open sourcing of Java, Blizzard said that the OLPC may move to Java as it is close to native speeds thanks to Java's jit (Just in Time) compiler and Python's interpreter being rather slow. One imagines that with the restricted hardware available that a slow interpreted language is the last thing you want, even if it is an exceedingly easy and powerful one. This is also the first impact I have seen from the open sourcing of Java.

From O(GL)LPC to O(W)LPC

An interesting story here:

Microsoft wants to make its Windows operating system available on the One Laptop per Child (OLPC) notebook computers, OLPC chairman Nicholas Negroponte said at the NetEvents conference in Hong Kong on Saturday.

...

"We put in an SD slot in the machine just for Bill. We didn't need it but the OLPC machines are at Microsoft right now, getting Windows put on them."

The SD slot is needed so that memory can be boosted sufficiently to run Windows. That probably won't be a problem in terms of cost, because memory just keeps on getting cheaper. But what's deeply ironic here is that the current price of the GNU/Linux-based OLPC system - around $140 - is utterly dwarfed by the cost of Windows. Obviously Microsoft will offer a cut-down, el cheapo version, but nonetheless the unjustifiable disparity between hardware and software costs is striking.

Microsoft's interest is understandable - it doesn't want to lose a potentially huge and impressionable market. What is less understandable is Negroponte's willingness to give up all his fine principles of empowering children, and to allow them to be shackled by closed source/DRM/Trusted Computing - for what looks like a rather pathetic and unbecoming reason:

"I have known [Microsoft chairman] Bill Gates his entire adult life. We talk, we meet one-on-one, we discuss this project," said Negroponte, according to a transcript provided to vnunet.com.

Gosh, you must be important. (Via Slashdot.)

Bicycle-Powered GNU/Linux System

I must confess my initial scepticism to the One Laptop Per Child project has waned a little, not least because it does seem to have some genuinely cool technology behind it.

But I was nonetheless intrigued to find that there is already something similar that is not just on the drawing board, but off it and deep in the field (or jungle/savannah as the case may be). It's called the Inveneo Communication System. It uses GNU/Linux (of course), and draws as little as 12 Watts of power, which can be supplied by sun, wind or bicycle.

One Laptop Per Child: Pedagogically Suspect?

I've been a little dismissive of the One Laptop Per Child (OLPC) project, but having read LWN.net's fascinating interview with Jim Getty, the Vice President of Software, I have become more impressed with what it is trying to do, and how it is trying to do it.

Unfortunately, it seems that the Indian Ministry of Education does not share my new-found enthusiasm. According to Heise Online:

The Indian Ministry of Human Resource Development (HRD), the country's ministry of education, has rejected the suggestion made by the country's Planning Commission to take part in the "One Laptop per Child" (OLPC) program. In a letter sent to the Planning Commission last month Education Secretary Sudeep Banerjee of the Ministry of Human Resource Development stated that the case for giving a computer to every single child was "pedagogically suspect." This the Indian daily The Hindu reports. "It may actually be detrimental to the growth of creative and analytical abilities of the child," the secretary added. Classrooms and teachers were more urgently needed than "fancy tools," the Education Secretary went on to say.

It's a fair point, and a concern that research will need to show is unfounded when the first deliveries of the laptops take place next year (assuming they do).

Microsoft is Right - No, Really

At first sight, the $100 laptop has everything going for it: it is based around open source software, uses renewable energy (you wind it up), and is trying to do something really worthwhile - put computing into the hands of children in developing nations.

But I have to say that, even though it is being done for all the usual wrong reasons, Bill Gates's alternative solution - to use a mobile 'phone to provide the processing power - seems spot on to me. As prices continue to plummet, mobile 'phones will soon be affordable even in countries with very low per capita incomes.

Moreover, today's mobiles are already computers: they play music, take digital photos, and often run office-type software (to say nothing of games). And they just keep on getting smaller and lighter. Convergence from the other end - putting a 'phone into a portable computer - does not lead to the same end-result for one simple reason: there is a limit to how small you can make a keyboard.

Mobiles get round this problem by ignoring it: keyboard entry is done either in a minimalist form (texting) or not at all. As I've written elsewhere, once voice recognition systems are good enough to cope with breathless speech on the move with significant background noises, nobody would even think of using a keyboard; typing will become some ancient art like thatching or dry stone walling.

Better, then, to work out ways of turning what will soon be the ubiquitous mobile into a teaching tool. Better still, if that tool were based on some form of GNU/Linux for mobiles rather than Microsoft's proprietary solutions. But I fear this is unlikely to happen: the MIT project has achieved a technical, economic and political momentum that means it will carry on regardless of whether it is actually the best solution.