Inside UP : ASUS

Look! Windows on the OLPC XO!

jamesu — Fri, 16 May 2008 02:00:03 GMT

Today Microsoft and the OLPC are announcing support for Windows on the OLPC XO computer. The two organizations will work together on several pilot programs in emerging market countries starting next month, and the offering will RTM in August or September. Initially it will only be available in emerging market countries where governments or NGOs are subsidizing the purchase of a large number of PCs for students, but there is the possibility of making this available for other customers through a broader set of channels at a later point in time.

From our perspective, Windows on the XO is a nice addition to the portfolio of products and services Microsoft has created to help transform education, one of the key themes of Unlimited Potential. It builds on the work we have been doing with partners like Intel and with programs like Partners in Learning, which has now reached over 100 million students worldwide.

And as you can see from this video featuring UPG's own Bohdan Raciborski, the Windows port to the XO is a snappy release that doesn't cut features or functionality in order to work in the constrained memory and storage environment of the XO.

It is the same basic Windows XP implementation that runs on the Intel Class Mate, ASUS eeePC, and other products in this emerging class of ultra low cost laptop PCs. As I have posted earlier, we had to write multiple custom drivers and a BIOS to get Windows to boot from an SD card in order to do the Windows port to the XO. This is the initial implementation customers will be able purchase when the product RTMs and will be a "Windows only" XO that Nicholas Negroponte himself has described as running "really fast." Customers can also choose to buy the existing Linux/Sugar XO. Longer term, the OLPC plans to write a new BIOS and increase the amount of flash storage on the XO to support a "Dual Boot" option that would enable children to use either Linux or Windows on the same machine. This is fine with us as long there continues to be an excellent Windows experience on the XO.

So you may ask, why is Microsoft doing this?

The answer is simple: people are asking for it, it transforms education and it leads to the creation of jobs and opportunity.

You can classify demand for Windows on the XO into three groups. The first group consists of people who have fallen in love with that cute little green laptop with its excellent industrial design but are committed to Windows. I wrote last fall about the guys from the Romanian Ministry of Education who like Windows (their teams regularly place in the Imagine Cup) and thought it would be cool to evaluate Windows on the XO. Another example is the NGO Save the Children, who are interested in sponsoring projects with the XO but as an IT organization have a ~~Windows-only~~ Windows-standard policy. Any extra money they spend in IT supporting multiple operating systems or technology camps is money diverted from their core mission around service, which for them is not a good thing.

The second group involves governments who are considering deployment of the XO en masse but also want the low deployment risk and broad support that the Windows ecosystem can provide them. Let's face it, there are hundreds of millions of Windows machines out there in the world today, which means there are thousands and thousands of people who know how to deploy, support, fix, and upgrade them. Despite the "let the kids fix their own computers" mindset that exists in some parts of the open source community, what we call at Microsoft the "IT Pro" is exactly the type of person that is needed for these large scale education deployments. As we all know, computers break, and asking children and teachers to fix them is not always the best solution. When I presented Unlimited Potential in Guatemala to a gathering of Ministry of Education types from across the region, the slide that generated the most interest was the one that described Microsoft's IT infrastructure optimization framework for large scale education deployments. Based on that customer feedback, we've decided to invest even more into a formalized national PC deployment methodology that we are starting to roll out right now. And believe it or not, it's easier to find Windows system administrators in places like India and Africa than it is to find Linux system administrators, and the Windows IT Pros cost less. We'll be releasing a study on this next month, so stay tuned.

The third group involves people -- usually policy makers -- in governments who see a direct link between technology investments in education and the need to expand the skills capacity of their workforce on a national scale. In other words, they want to implement policies that can positively impact education and set the stage for better employment opportunities for their citizens. They see Windows as a key ingredient for making this happen because it is the software environment used by so many businesses around the world.

Microsoft has created the Unlimited Potential initiative around the themes of transforming education, fostering local innovation, and enabling jobs and opportunity. Today's announcement gives us the opportunity to reinforce how these three themes can support each other given the right scenario and the right set of tools. If we can provide children with a great learning experience, and do so in a manner that involves a massive scale with the right level of (local) support, it has the potential for being transformational across multiple fronts. It's pretty exciting.

OLPC in the News (Part 2)

jamesu — Thu, 06 Dec 2007 01:17:17 GMT

Well, I will be flying out to Cambridge next week for my first meeting with some of the people at the OLPC, and I have to say I am looking forward to it. Some of my UPG co-workers from Microsoft have been meeting with the OLPC team for about a year now, but since I am a relative newcomer to our group, this will be my first trip.

One of the things we will be discussing is the status of our port of Windows XP to the OLPC XO computer. There have been suggestions in the press by Nicholas Negroponte and others that “Windows already runs on the XO.” That’s not really the case yet, and with the attention the OLPC’s “Give One Get One” campaign is getting, along with the strong level of interest we are receiving from some Ministries of Education and NGOs in buying a version of Windows for the XO, we thought it would be useful to provide some clarity on the topic.

For starters, we are hard at work on the project here, and we are using an approach that is a little unusual for Microsoft in that we are managing the entire process of adapting and testing an existing version of Windows for a new PC. Usually the hardware vendor does this. And the Windows port to the XO is by no means done. Between Microsoft employees and third party contractors that we have brought into the effort, we have over 40 engineers working full-time on the port. We started the project around the beginning of the year and think it will be mid-2008 at the earliest before we could have a production-quality release.

Because of this, we have not announced formal plans to support the XO yet, and we will not do so until after we start getting feedback from our first limited field trials starting in January before we make the final call. We do not want to set expectations we subsequently cannot meet, especially when it comes to supporting the children’s machine. For governments in emerging markets evaluating purchases of Windows for the XO, this means that so far we are not announcing an availability date, pricing, or support policies. In fact, you should not yet assume that Windows on the XO is a done deal. We are hopeful that we will have a different story for you within six months.

It also means that if you are in the US and Canada and are participating in the “Give One Get One” program, you need to understand that Microsoft is not currently planning to support a retail consumer release of Windows XP on your XO computer.

Why is this work taking so long?

Flash
First, the XO computer uses flash memory instead of a hard disk drive for storage. This is one of the reasons OLPC can get the production cost of the computer down to $188. This is a relatively new class of machine, and we have to do design work to get Windows and Office to work reliably and with good performance using only 2 GB of storage. The XO actually only comes with 1GB of flash, and we asked the OLPC to add a slot for an internal SD card that will provide the 2 GB of extra memory needed to run our software. (By comparison, an entry level $499 Dell laptop comes with 60 GB of hard disk storage.) The potential payoff for students and schools from this work, of course, is that the tens of thousands of existing educational applications written for Windows can potentially run on the XO. As part of this engineering effort, we have to design a new BIOS – the layer of software that runs between the hardware and an operating system -- to have Windows boot and run off the SD card. For us this is new work and requires a design and processes for supporting the XO’s custom SD interface and for the installation of Windows on the SD card, both at the Quanta factory that manufactures the XO hardware and also in the field.

For much of this XO flash design, we are able to leverage the work we did to get Windows to support the Intel Classmate PC, another computer that uses flash memory for storage. However, the Intel computer comes with 2GB of flash storage, so we did not have to use the SD card approach we are designing for the XO. The Classmate port took us about 9 months, but we started that effort a year and a half ago. A third example of these low cost “Flash PCs” on the market is the ASUS Eee PC, and surprisingly enough getting Windows running on this computer required a significantly shorter amount of time because ASUS used a more standardized approach to its hardware design compared to the XO. In technical terms, ASUS put the flash drive behind the IDE disk controller, making the flash storage "look like" a hard disk drive to Windows.

Microsoft plans to publish some formal design guidelines early next year that will help Flash PC manufacturers benefit from our early work so they can design machines that enable a great Windows experience at as low a cost as possible, and with a minimum of custom design work necessary to get Windows to run on their machines, such as we have encountered with the XO.

Cool New Features
Secondly, as we all know there are many innovative features in the XO computer that set it apart from other designs, and we are working with partners to write the driver software so that Windows can support all of them. This includes drivers for the XO’s wireless networking, camera, graphics processor, audio system, and the various user input devices (game pad, writing pad, touch pad, directional pad, and mouse pad.) There are ten custom drivers in all that we are writing. We also hope to support the XO’s mesh network design, its power-saving “e-book” mode, and its capability for excellent screen visibility in full daylight.

And we have a different support model than OLPC is envisioning: we are not expecting K-6 school children to access the source code and do their own programming in the event they have to fix a problem in the computer. Certainly, we think there is a role for students in the support of school computers -- in fact, as part of our Partners in Learning program we have trained over a million kids in a student helpdesk program (like in this case study from Brazil) -- but we also think that local entrepreneurs and businesses need to play an important role here when you are talking about deployments involving tens of thousands of computers.

We want to support these new XO features without sacrificing compatibility with existing Windows applications, and we want to deliver an out-of-the-box user experience similar to the quality people expect from Windows running on more expensive classes of machines. All of this takes a lot of work.

Fast Moving Partner
Finally, we are doing this engineering work for a moving target. It is literally like designing parts of a car – well, actually a school bus -- while it is running down the highway at a high speed. I am not meaning this as a knock on the OLPC organization, because they are a small group of people doing an amazing amount of innovative design work in a short period of time. But we have only received a handful of machines for most of the last year, and the XO team was doing some hardware design changes as recently as this past August. This affects our schedule.

Much of the technology in the XO is developed using open source technology licenses that make it difficult for engineers employed by commercial software companies like Microsoft to work directly on the project. For this reason, we also had to follow a complicated process to figure out interfaces for many of the XO’s hardware components and to deal with some of the hardware bugs they were reporting in their design process in order to make progress on our port. All of this slows us down, but that’s OK given our overall shared mission here.

We appreciate the support we are getting from the OLPC team, and we know the focus their engineers need to get the XO out the door and into the hands of students. Now that they are finally shipping, our ability to support the XO with a quality release of Windows is accelerating. I also have to say that if our team continues down the path they are on and the system performs as we hope, then that cute little machine with the Wi-Fi ears will run Windows!

What Does This Mean for Users?
The Unlimited Potential Group at Microsoft is developing technology to enable social and economic opportunity for “the next five billion," and one of our key focus areas for doing so is through the transformation of education. As part of this, we are investing in programs and partners around the world to foster innovative schools, innovative teachers, and innovative students. We have a lot going on here, and there is clearly a role for low cost hardware as part of this vision. In fact, there is a good alignment between what OLPC is trying to do and what we are trying to do. And frankly, nothing would please us more than seeing hundreds of thousands of these XO computers that are now starting to be deployed all running Windows given the very high interest that has been expressed in the market for it. We are committed to developing a quality port of Windows XP for the OLPC XO computer, but we still have a lot of work to do to complete the effort.

OLPC in the News ...

jamesu — Wed, 28 Nov 2007 20:09:22 GMT

Well, it seems like the OLPC organization is in the news a lot lately, even more so than usual. Stories that caught my eye over the last week included a CNN report, a BBC story from Monday, and of course the Wall Street Journal page A1 story that came out this last Saturday.

Microsoft's approach to the OLPC continues to be that we will work with them to see if we can get Windows to run on the XO machine -- there is still a lot of technical work to do, despite what you might hear in the press -- but otherwise we need to remain focused on our Unlimited Potential mission for enabling social and economic opportunity for the next five billion through transforming education, fostering local innovation, and enabling jobs and opportunity. We are applying a great deal of energy across these three areas in pilot projects around the world, and we don't really want to get distracted by the public rhetoric taking place around the OLPC and their XO machine. We are working with partners on a broad spectrum of solutions for education in emerging markets -- and low cost computing is just one of them -- but we also have pilot projects in other areas ranging from rural kiosks to new approaches for subscription computing, new applications for cell phones, new models for Internet cafes/community centers, and new approaches for mobile and remote access to the Internet. We have a lot going on and really need to focus first and foremost on the needs of the communities we are serving. Our mantra in all of this is "Relevance, Access, and Affordability."

The WSJ article was kind of cool in the sense that I was interviewed and referenced in the story, even though I didn't land a direct quote. I've been at Microsoft since 1995, and this was the first time I have appeared on the front page of the Journal. (By the way, my sister Lisa -- who also works here -- was featured in a page one WSJ story last year, for those of you who are keeping tabs on the Utzschneiders.)

Anyway, I've thought a lot about this whole OLPC phenomenon, and the best way for me to summarize my thoughts on the topic is to refer you to two quotes, both from bosses I've had at Microsoft.

The first is from David Vaskevitch, one of Microsoft's CTOs. He was an early mentor of my career here, and at one point I ran a technology incubation team working for him. David always liked to remind me that "the technology industry consistently overestimates what it can accomplish in 2 years, and consistently underestimates what it can accomplish in 10". This is coming from a guy who chose to center his 1996 Professional Developers Conference keynote around the emerging importance of digital photography -- we all thought at the time that he was nuts -- but look at what happened 10 years later. It's now one of the most widely used scenarios on the PC today (and among other things, a staple ingredient for how I create this blog.) And for what it's worth, I used this quote in my interview with the WSJ to summarize our view of what Nicholas Negroponte and the OLPC are doing.

The second quote comes from Doug Burgum, the man who spent 25 years building the Great Plains/MBS business into what ultimately became a billion dollar division for Microsoft before he retired this past summer. Doug had an amazing capacity to inspire a community of channel partners into creating an ecosystem around a shared vision and more importantly a shared set of values. His quote -- it actually originated from Margaret Mead, but Doug liked to use it a lot -- was to "Never doubt that a small, group of thoughtful, committed citizens can change the world. Indeed, it is the only thing that ever has." One of the powerful ideas behind the OLPC is their approach for harnessing the power and excitement of a community to accomplish a shared (and in this case, noble) mission. We know at Microsoft what this can feel like; sometimes people forget that my company has a lot of experience with building communities organically. There's nothing like the feeling you get when you start a parade!

So my view on the OLPC is that Nicholas, Walter, Mary Lou-- all people I've never met but whom I admire at a distance -- are a group of dangerous dreamers (another Dougism) who are out to change the world and could have a huge impact on education over the course of the next ten years, but not so much in the next two. I love the boldness of their vision, their focus on serving the needs of poor children, and their desire to do great things.

But I also know the reality of the physics of the IT industry and the difficulty in trying to go from zero to millions of deployed, functioning, supported machines in a matter of months. About the nature of how this industry works, where one group may come up with an idea and then other organizations or individuals build on the idea and come in from seemingly nowhere (hello ASUS!) with a different type of solution to fill a vacuum created by the original vision. (Ask me how I felt after I read the first public draft of the Enterprise Java Beans spec, a document that was "inspired" by work we were doing on COM and MTS in the mid-90's.) And how the implementation of IT visions ultimately comes down to customer choice, because people -- even people who work in government Ministries of Education -- are rational actors who select things that are in their best interest and take into account price, roadmap, TCO, pedagogies, politics, local infrastructure, support, bake-off results, the need for measurable outcomes, you name it ... the whole variety of factors that go into a complex government purchase process.

It takes a village to buy a computer, and it's always harder than you think it will be.

But that's all OK, because the OLPC vision isn't going to go away. There will be a permanent role for low cost, flash-based PCs in national education and technology policies. The XO will survive and evolve, and I bet every laptop vendor on the planet including Dell and HP will have a competing machine within 24 months. A new ecosystem of collaborative, social network-inspired and Internet-enabled education software will emerge. Cell phones will play a bigger role in this space than even Nicholas is publicly acknowledging. And kids and teachers will author a lot of the content.

Dangerous dreamers who assume they will change the world in two years but actually do so in ten, in a manner they never initially anticipated. That's my personal view of what the people at OLPC are trying to do. I love the industrial design, I love the screen, and I love the rabbit ears. I wish the team well. But there are other dangerous dreamers out there, and ultimately it will be the magic of software delivered in a sustainable manner that will be the key to transforming education.

But now I need to go back to work.

Buchalost

jamesu — Thu, 18 Oct 2007 09:34:41 GMT

October 17, 2007

I began the morning in Bucharest today by going for a run and getting terribly and completely lost. I mean 45 minutes of "Do you speak English -- do you know where the Howard Johnson's Hotel is?" lost. The team ended the day having to cross a highway and walk along the side of the road a ways to get to our car. But in between these two we had a fabulous day, learning a lot about some of the progressive steps Romania is taking to apply ICT in a sensible way to improve education in their country. And as you can see, the weather was way nicer than Moscow.

We started the day (post lost run) in the Microsoft office where we met with people from Romania's Ministry of Education. Included in the meeting was Professor Adrian Petrescu, who is considered sort of the godfather of technical education in his country. Here is a photo of Dr. Petrescu along with Silviu Hotaran, Microsoft's country manager for Romania. Silviu (on the right) was a student of Dr. Petrescu's in the university.

Also attending from the Romanian government was Catalin Grosu and Decebal Popescu. (Yes indeed, there is a person in Romania who looks just like Jack White from the band the White Stripes and is named "Decebal".) Decebal seemed very smart and was passionate about IT curricula. Among other things, he asked about our plans for porting Windows to the OLPC XO device. Will Poole gave a good answer, the shorthand version is "Because the device is exotic and requires so many custom drivers, it's hard. But we are trying."

Before the 1989 revolution here, Dr. Petrescu designed and deployed across the country a microprocessor based system that used a TV set for display, a cassette recorder for storage, and was based on the Sinclair. Needless to say, he has a lot of experience with national PC programs. He also says he is inspired both by BillG's book "Business at the Speed of Thought" and Nicholas Negroponte's evangelism efforts around 1-to-1 computing, namely, that the best educational technology model for a country involves students with their own PCs with roaming access to the Internet. But he also clearly believes that there is no single optimal technology solution to this problem -- this is a theme I am hearing every time I go on the road now -- in this case because kids from wealthier families want machines with more capabilities than the Asus or the OLPC XO, but this class of machine might be ideal for rural students from lower income families. Early results from one of the pilots in Romania indicate that kids get quickly bored if there is limited software on one of these low end laptops.

Apparently Romania's Prime Minister grew up using one of the Sinclair-like machines, has a technical background, and did a side-buy-side demo of the OLPC XO and the Intel ClassMate PC to members of parliament, explaining how each of the machines worked. That would have been fun to watch. Also, Decebal was quite enthusiastic about the Imagine Cup (Romania won it last year!) and was effusive in his praise for the support they received from the local Microsoft team. Finally, we walked them through a slide deck describing our education strategy, and they expressed interest in learning more about Multipoint and about Microsoft's infrastructure optimization model for education. We will definitely follow up.

From there we went to visit Siveco, one of the largest software ISVs in Romania. We met with Florian Ciolacu, Florin Ilia, and a partner account manager (Flora?) who all spent an hour walking us some of the impressive work their company has done in education. Their product is called AeL, a suite of interactive K-12 learning content consisting of 1,800 interactive modules (called "objects") built on the Microsoft stack. They even referred to it as DHTML, not AJAX. You can find samples of their work on the Innovative Teachers Network, including this self-paced chemistry class. (Apologies, it's in Romanian). Florin says that about half of their classes are translated to English, and about 300 to Russian.

As far as we can tell, Siveco is one of the few technology partners who have successfully implemented a national education program. Over the last 6 years, in partnership with the Romanian Ministry of Education, they have deployed to Romanian schools 76,000 Windows/Office desktop systems and 1,500 servers; their software was used as the basis for training 80,000 teachers and 3 million students. The company is expanding to the CIS countries, Cypress, and the Middle East. They gave me a copy of AeL and I plan to start demoing it on my travels.

After Siveco, we went to visit RTC, a conglomerate of 70 companies that is partnering with us on a family education PC project in Romania called PC@casă. This is an incubation where we partner with an OEM, retailers, and software partners to package up a low cost PC loaded with education software. We met with RTC's leadership team and did a fun interview with a family that purchased one of these systems (the dad Nicolae and his son Alexandru are shown here, mom Camelia was at work). We'll be showing a video we took of this family at our press event in Budapest tomorrow. I will try to get my hands on a copy and post it.

Finally, we had a late lunch with Varujan Pambuccian from the Romanian Parliament. The guy is awesome. He has been an elected member of parliament for 12 years and is the chairman of their IC&T Commission, and he used to design computer systems at the country's Technical Institute (including compilers, and is show here with his fellow former compiler designer Paula Apreutesei, Microsoft's citizenship lead for Romania). We all had a great back an forth discussion, ranging from voucher models for the national "Euro200" PC program to motherboard design to the future of networking in countries in Romania. I loved the fact that a guy who is still hands-on technical (he is working on some patents for TCP/IP extensions) is engaged as an elected official in setting ICT policy for his country.

The meal was in a "classic" local restaurant called Casa Romaneasca where they served us a wonderful platter of meat, family style, which in turn resulted in "meat comas" hitting us on the plane to Budapest 90 minutes later. (Last photo courtesy of Will Poole).

More tomorrow from our launch event...