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?
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.