OLPC myths

Myths

Separated by "False" and "Falsehood" (inaccuracies) misunderstandings, and other responses where appropriate.

The laptop does not have internal storage such as a hard drive.

False: The laptop has 1GB of internal Flash memory similar to the inexpensive thumb drives sold at many computer stores. Operating systems can be installed and/or files can be saved on this memory. The laptop also has USB ports for external hard drives; so internal Flash storage can be used for the OS and some file storage, and common external USB drives can take up the slack if needed. Further, the laptop has an SD-card slot for further expansion.

The laptop will be really clunky with a hand crank on the side

Early prototypes included a hand crank, but it was removed in subsequent versions. The actual shipping units will use an off-board human-power system, connected to the power brick. Candidates include a foot-pedal charger similar to the Freecharge portable charger, solar panels, a crank, and a pulley system.

You're expecting this to be a magic bullet for poverty.

False: Not at all. It is simply a tool for education and communication and only helps, in part, in contributing to the entirety of aid programs where these laptops are distributed. Nevertheless it provides access to education, health, technology, economic opportunity, and more, and a few children will be able pull themselves out of poverty with no other assistance.

The laptop isn't powerful enough to run modern 3D games and other resource-heavy programs such as video-editing software.

True in some cases but irrelevant: That's not the purpose of this laptop. It is designed to be an inexpensive way for people of limited means to use a computer for such things as internet and educational software. The choice is not currently between this system and a more capable one: it is between this and nothing. This is better. [What programs the laptop will run well is more about how well the program is written than the functionality of the program, e.g., the laptop will have little forgiveness for programs with memory leaks.]

The laptop will be Linux-based

It is true that the first prototypes will run a slimmed down version of Fedora Core with the Sugar user interface produced by Red Hat, however other systems have also been considered and could be loaded later.

The laptop will run a Microsoft Windows operating system

• True: Microsoft is working on a Windows based system that can be executed on the XO laptop with substantial extra storage.

• False: There is no strategy change. The OLPC is continuing to develop a Linux-based software set for the laptop in conjunction with Red Hat. But since the OLPC project is open we cannot (and maybe even don't want to) stop other people from developing and supplying alternate software packages.

An old Pentium laptop can do the same thing.

False: The point of this laptop is to keep people connected with the modern computer net-based society. Using a laptop that may be on its way to obsolescence from a second-hand store, or building new expensive Pentium laptops for this purpose isn't feasible. You have to design something specifically to answer all the requirements of the XO laptop. If we could make a reliable $2 laptop that is modern and can do everything required of it in our program, we would absolutely make such a device. Another problem with the "old- or used-computer" approach is that it doesn't scale. The overhead of deployment and support would overshadow any potential economies in terms of the capital costs. A final, insurmountable problem with the "old- or used-computer" approach is power. The XO laptop uses an order of magnitude less power than the typical laptop. It is both environmentally reckless and economically infeasible to power used computers in a developing world.

You're forcing this on poverty stricken areas that need food, water and housing rather than a laptop.

False: Not at all. Like it was said earlier, this is only a tool and should not be seen as more than that. We agree that other more urgent matters must be attended to before you insert high tech into the situation of poverty.

Not everybody agrees with that idea. Some think that access to the Net is the fastest way for poor people to get the political clout to require their governments to provide services to them. Or to get the education for real jobs that take them out of poverty completely. Or access to innovative technologies for providing food, water, clothing, shelter, energy, etc.

But we believe education and communication with the modern world to be important as well. Food, water, clothing and other necessities come first. Nevertheless, a world view and good education can do wonders for a child's mind and continued health. Computers, especially those that are networked, have shown to be development 'multipliers', that is they help to improve the delivery of medical, educational and communication services.

Whilst I applaud the innovative nature of the project and particularly the conceptual model surrounding "Sugar", it worries me that it is being foisted upon the so-called third-world communities; it all smacks too much of the flavour of some kind of social pedagogical "experiment". There is, I suppose, a great deal of sense in the argument that such communities have not been exposed to and pre-conditioned by existing interfaces and their established standards, and will thus be more amenable to alternatives. However, I can't help feeling that strategies for optimum learning are by their very nature culturally dependent, as indeed are any notions of what might be considered valuable knowledge. I can't help thinking that it is unlikely that these "third-world" communities have had much genuine say in the OLPC project's design and planning... Once again, "wisdom" seems to be in the hands of the rich, generous, patriarchal white nations of the "West". Perhaps the "experiment" would have been better carried out in the US?

Nevertheless, I find the concepts truly fascinating and regardless of my concerns I certainly hope the project proves successful. All going well, the interface will hopefully be flexible enough to mould itself to the cultural behaviour and activities of its users. The hierarchical model of the traditional desktop GUI, on the other hand, certainly models the Linnaean-like structures of the colonial past and, I fear, of the postcolonial present.

Mcewanw 19:26, 13 September 2007 (EDT)

A few comments about "pedagogical experiments" and Sugar: OLPC pedagogy is based upon Constructionism, the gist of which is that you learn through doing, so if you want more learning, you want more doing. While this approach is not epistemologically agnostic, it is for the most part culturally agnostic: it is—by design—amenable to adaptation to local cultural values in regard to what "doing" is appropriate. Projects tend to be grounded in local contexts; one role of the teacher is to help the children shape their constructions within these contexts. Constructionism is not a new idea—it has been used by teachers and learning for many decades, not just in the "West" but in virtually every corner of planet. Many of its greatest theorists and practitioners come from the developing world.

The role of Sugar is simply to provide some affordances that enable children to explore, express, and communicate. Sugar is a community project that has contributors from a diverse base; feedback from teachers and students in roughly one-dozen trials in the developing world has greatly influenced the design. Further, it is—by design—free and open. --Walter 04:11, 15 September 2007 (EDT)

The debate that weighs the benefits of computer technology and constructionist pedagogy against the benefits of improved access to food, water and electricity in less developed countries fails to take note of one clear (if not immediate) financial benefit of distributing cheap laptops to students: savings on and improved access to textbooks, literature and libraries. In the United States, a new textbook will cost anywhere from $30 to $150, depending on the subject matter. It may be reasonable to stipulate that in poorer countries textbooks would cost less, but they may even cost more due to difficulties in distribution; regardless, in many poor classrooms there are fewer textbooks than there are students, and textbooks that do exist are old and out-of date. Furthermore, the multi-million-dollar libraries that are common in American schools, towns and cities are almost non-existent in the poor rural areas where the XO laptop is intended to go.

A simple calculation can place the cost of the XO laptop in context. Assuming that a $20 textbook will last five years and be used by five students in its lifetime, and assuming that each student uses five $20 textbooks each year, each pertaining to a different subject (math, science, english, french, and history, for example), a group of five students can be said to use $500 in text-books over five years. By this calculation, at a $100 price-point, the XO laptop (with a projected lifetime of 5 years), loaded with up-to-date open-source textbooks, is cost equivalent to providing physical textbooks to a student for five years. Yes, it is true that the XO laptop is priced at $180, that printed textbooks can physically last longer than five years (if not staying up-to-date), and that as of yet open-source textbooks are rather primitive. Students need more that just textbooks to learn, however: they need access to literature, to primary research materials, to encyclopedias, to tutorials (which can be provided by access to internet resources like gutenberg.org, wikipedia.org, and Google books). They also need work-space (notebooks, pen and paper) and light to read at night (which can be provided by a mechanically charged laptop). And with increased production levels, the price of the XO can be expected to go down.

It is difficult to argue that education is not a necessary component to poverty reduction, probably being more effective than food donations or development aid; it is even more difficult to argue that children can be taught without books. Although the price-point of the XO laptop does not make it an obvious replacement for the bare-minimum expenditure on textbooks that must be made to minimally educate a child, when seen as an opportunity to provide poor children with the same literary resources as are provided to American children--and without making the same investment in physical libraries--an investment in laptops starts to seem very reasonable. Furthermore, laptops are the best solution to certain problems that are unique to poor rural areas: limited means of distribution for bulky textbooks, a lack of night light for reading and study, insufficient capital to create large redundant libraries in multiple locations, and decentralized communities that make the use of any library difficult (and possibly unsafe) due to travel distances. The infrastructure investments that are required in order to make wireless laptops maximally effective--specifically, investments in wireless infrastructure--have a multiplicative economic effect that is not limited to the domain of education.

Leaving aside the more innovative pedagogical aspects of the OLPC project, the distribution of laptops solves basic problems that are inherent obstacles to educating children in less developed countries: most importantly, access to essential reading materials. In fact, the broad distribution of laptops (along with the development of open-source textbooks) is probably the only practical way to give rural children in poor countries access to the same quality of primary and secondary education as children in rich countries at a reasonable cost. --Legutierr 06:13, 21 December 2007 (EST)

This isn't a laptop. This is a gadget.

Falsehood: What does this mean? To call it a "gadget" implies that this is nothing more than a toy or an insignificant object of interest. Can a commercially available PDA do better? Highly unlikely. Can a cell phone do better? Why waste a cell phone that may not have the needed features in an effort to avoid designing something to really solve these problems? This is not to say that others can't come up with better ideas; we encourage it. In the strictest utilitarian sense our hope is that this technology we call a laptop can do much more than mere gadgetry. And we're confident it can.

You can't use a laptop in a place that might not have power.

False: If the laptop comes with a method of inexpensive self-contained rechargeable power, such as wind-up power that lasts a good long time, this is not true. [We are aiming for a minimum of a 10:1 ratio between time put into reading the eBook and time human-powering, i.e., one minute of cranking gives you at least ten minutes of reading.] You might be surprised at the number and variety of energy sources available in poor countries, including solar, wind, water, bicycle, animal, biomass, etc.

It is not made of recyclable components

False: The XO is the greenest laptop ever made. EPEAT (www.epeat.net) is an organization that measures the environmental impact of laptops. OLPC is in process of applying for a rating for them, which we believe will be excellent. XO appears destined to be the first laptop to receive their Gold Rating, and in fact, it has been suggested that the XO may warrant establishing a new, even higher rating.

It will contribute to the landfills worldwide if made in vast quantities

We hope that is not true. If the program is run well by governments all or most laptops will be accounted for. No doubt some laptops will find their way into trash bins and garbage dumps, where there are strong financial incentives for almost all of them to be reclaimed. Will broken laptops be brought back and recycled or parted out for new laptops (refurbished)? Hopefully. And hopefully distribution and reclamation will be conducted in a responsible manner.

If others are coming up with their own ideas then there must be something inherently wrong with your idea.

Falsehood: Not so. Like mentioned earlier, if other groups and businesses come up with their own ideas then we encourage it. That doesn't automatically put a value judgement on competing ideas. Some ideas or projects may have strengths and weaknesses that others do not. We can only learn from each other to better each other's ideas and we hope we will in the spirit of goodwill towards those who need it. But what we hope this does not turn into is petty rivalry and cutthroat politics in business that is not conducive to a cause we happen to champion.

This laptop will do more harm than good.

How so?: If it's simply a tool as water purifying machines are tools, how will it harm the people it's intended to help? If you're talking about the digital divide in most places where this program may be instituted, think on that a minute. If this laptop does what it's intended to do it can only open avenues to better close that divide. If you're talking about the environment, read the entries on that further up the list. Or do you think that these laptops will, for example, destroy languages and cultures? Hardly. Even now, minority languages that seemed to be dying out, such as Hawai'ian, Welsh and Irish, and Yiddish are coming back. Why wouldn't that work for languages of Africa or Asia? These laptops will provide unequalled opportunities for saving languages and other endangered aspects of civilization and bring them to the attention and consideration of the entire world community. This is what we generally mean by "communication" and this laptop can only help, not hinder, in the achievement of this goal.

Why this hasn't been done before is because there must be a very good reason against it.

False: All things have a beginning. And not doing things just because others before you haven't done those things is no reason not to do them or make excuses why not to do them. Most likely because possible previous attempts have failed (presumably) is because the right technology just wasn't there to begin with. Now we have technology that is cheap enough and available enough to attempt something of this magnitude. That's how it's always been. We are trying to "stand on the shoulders of the giants who came before us" and learn from their mistakes, and we have had to invent some things from scratch to bypass some of the well-known pitfalls of such a project. This evolution of technology is based on Moore's Law. The evolution of the epistemology--recall that this is a learning project--is based on 40 years of research into technology and learning. Eventually the written word spread across the globe and obviously was developed as civilization developed. It had to start somewhere with someone. Same with technology and its eventual seeming ubiquity.

OLPC has no plan for (insert topic of objection)

People say that OLPC has no plan for recycling the laptops, or training teachers, or getting software into local languages, or preventing wholesale theft and resale of the machines, or a host of other things that we clearly should plan for. The fact is that it is too soon to have an announced plan for any of these things. But lack of an announced plan does not equate to lack of planning. Significant numbers of people are putting their best thoughts and other efforts into these problems, and will have much to say at the appropriate times.

On the other hand, how can we plan in any detail for such huge transforming events as bringing a generation out of poverty? Look at countries where it has happened, like South Korea, or is happening, like Thailand. Who could have predicted twenty or thirty years ago where they would be today, and what they would need next? Who could have predicted that South Korea would become the most highly digitally-connected nation on Earth, or the state of the North Korean and Burmese refugee problems?

So the kind of planning we have to do is what software developers call Agile Planning. We have to know what we can do next, and we have to create a process for understanding what happens when we do it, and how we can proceed from wherever we get to. The opposite of the infamous Soviet or Indian Five-Year Plan, or the Waterfall model of software development, where everything is supposed to be known in advance, whether it can be or not.

What do we need to do next?

• Build and test the computer, and get it ready for production
• Get more software for it in more languages
• Get financial commitments for the first production run and field trials
• Plan enough of the training and logistics for the trials
• Research the trials
• Plan the next larger rollout

We have a pretty good idea how to do the first two, Sales & Marketing (AKA Prof. Negroponte) is working the third about as well as he can, and it's still too soon to do more than outline the last three. The problems of training and logistics will be different in every country. We will need to focus considerable energy on the issues that actually arise, and not wish for a plan that could meet every possible contingency.

How can we tell what happened?

• Pay attention, AKA research, done by professional researchers and by the people concerned. Read the children's blogs, for one major thing.

Then what?

• Oh, just another 200 or so countries where 6,000+ languages are spoken, major health problems, the odd civil war or tyranny, a little of this and a little of that, you know. No shortage of challenges. The perfect setting for a flowering of ingenuity that will put the Industrial Revolution to shame. I'm counting on the brainpower and energy of a few hundred million hungry children. You and I can't outthink them, especially not in advance.

So are you going to stand there cursing the darkness, or teach people to make candles?

OLPC is rewriting practically all code above the OS and UI gadget level in Python. On an underpowered platform, this will lead to intolerable performance.

This is probably false. The reasons probably include significant code not rewritten in Python (examples?), the fact that python isn't as slow as interpreted languages used to be (statistics?), and empirical evidence that things are working fine on the OLPC (examples?).

(When this answer is rewritten with specifics, please remove my signature: Homunq 15:30, 29 July 2007 (EDT))

The L2 Cache on the LX has a big impact on Python performance as well... --Walter 20:20, 29 July 2007 (EDT)

I also wonder how cpu intensive it is; I can't imagine it being lightweight in terms of the resources it needs.

Mcewanw 19:13, 13 September 2007 (EDT)

Python is in fact quite difficult to optimize. The language is very dynamic. Doing a good JIT is quite difficult because various things would frequently need to be checked for changes that might -- but probably won't -- happen. The normal interpreter does not JIT, which is quite a different situation from things like Java, C#, ActionScript, and JavaScript. In some cases we do in fact see intolerable performance. In other cases performance is merely OK. So actually this is not a myth. :-( AlbertCahalan 01:15, 18 September 2007 (EDT)

But there is a JIT (well, the Python equivalent thereof): Psyco. I understand that this increases memory overhead, a premium on the OLPC, but with copy-on-write Psyco itself need only be loaded once, and judicious use of optimization (only in the "intolerable" cases) should keep memory increases to a minimum. (Psyco is used by YouTube, which suggests it is not flaky academic code) --Homunq 01:15, 28 September 2007 (EDT)

As of November 2008, software build 8.2-767 demonstrates quite acceptable performance for a multitude of tasks including ebook reading, web surfing, email, and text editing. Education has flexible demands; if one kind of software runs too slow, the users can just try something else. I myself can choose between my XO and a 2GHz dual-processor desktop with a 19" LCD display. I choose to use the XO more than half the time because it is fast enough, plus is silent and portable. The XO does not do video very well (yet?), but I learn more from reading than watching videos. Thomaswamm 10:10, 7 November 2008 (UTC)

The price is rising horribly, it's already $189

It is depressing to see the price of the laptop going up every week. Isn't this going to make it harder to distribute as many laptops as we want to?

The price rises quoted in the press are based on converting the Taiwanese manufacturer's cost into US Dollars. However, this assumes that the payments for the laptop are sent to the US first and then to Taiwan. Since a large part of this price rise is based on the weakening of the U.S. Dollar, there are two ways that buying countries can mitigate the situation and maintain a lower price for the laptop.

First, they can send their money directly to Taiwan. Chances are that the relative values of currency between Taiwan and a developing country are not shifting that much. Secondly, they can take advantage of the fact that the US dollar is weakening which allows them to buy more US Dollars with their home currency than they could previously. If the US Dollar strengthens, then it is possible to lock in a savings by buying USD at its weakest and then paying the USD price later, after the price has dropped.

Bottom line is that you should not worry about price changes due to currency fluctuations. In addition, the current price of the laptop in the early days, is not terribly relevant. The really important price issue is how low we can push the price by ramping up the volume of laptops manufactured and shipped.

--Memracom 04:09, 17 September 2007 (EDT)

The pricing issue is quite complicated on the one hand—shipping, extended warranties, exchange rates, etc.—and quite simple on the other hand—OLPC always sells laptops at cost and has a mandate to drive the cost down over time.

The price has been hard to predict in that we've made some changes in response to feedback from the community, such as the upgrade to the Geode LX, and we are subject to trends outside of our immediate control, such as fluctuations in exchange rates, etc. Once we hit volume production, the price will come down. But as a non-profit, OLPC cannot do "forward pricing" where we sell laptops at less than cost with the expectation that future profits will make up the difference—we have no profit—price always equals cost. The generous members of the OLPC association have subsidized the cost the laptop development so that no design or engineering costs are passed to the children. Our foundation is trying to raise money that can be applied to helping launch one-laptop-per-child projects in the least-developed countries.

An additional factor that is causing confusion about the pricing in the mainstream press and the blog-sphere is the difference between ex works pricing and the cost to get the laptop into the hand of a child, which may include shipping, warehousing, customs duties, local transportation, etc. These additional costs vary from place to place and deployment to deployment.

Yet another point of confusion is the pricing of the "Give One Get One" and "Give Many" programs. Both of these programs distribute laptops through the OLPC Foundation and include both "fair value pricing" and a built-in donation to help subsidize the cost of laptops to enable the foundation to reach out to children who would not otherwise be able to afford a laptop, even at $100. --Walter 08:12, 30 October 2007 (EDT)

Mass production has been delayed. You'll never be able to build the laptop.

The proof is now in the pudding...

It should also be noted that for most products, delays are not evident because the development process is kept behind closed doors. OLPC has been open about all aspects of the project in order to elicit feedback from the development community and the communities we hope to serve. We make no attempt to hide our "warts and blemishes" as we think that this is the most efficient way to discover (and correct) problems.

It is the exception rather than the rule that there are no delays in mass production for products as complex as a new laptop computer. That said, we have already begun mass production; thousands of machines are now being manufactured and shipped everyday.

--Jim 11:03, 10 December 2007 (EST)

The laptop is new so we don't know whether or not one laptop per child will work

While it is true that the XO laptop is new, the design process has been an iterative one—we have field tested almost 7000 prototypes before starting mass production. Certainly building laptops is not new to Quanta, the world's largest manufacturer of laptops. So while there is always some risk associated with any new product, we have every confidence that we have already discovered and corrected the vast majority of bugs.

But the "myth" regarding the newness of the laptop is really smokescreen for those who want to suggest that giving computers to children is a new and untested idea. OLPC is grounded in almost 50 years of research on how technology (computing in particular) impacts children's learning. We as a team have been engaged in field studies around the world for almost three decades, and we are just one of numerous long-standing research and development efforts. There is nothing new about the fundamental epistemology of "you learn through doing" and there is nothing new about giving children access to computing. What is new is the scope: we want every child and teacher to have the same opportunities that our own children have: access to knowledge through books, libraries, rich media, etc. and the ability to appropriate that knowledge.

The "Give One Get One" and "Give Many" programs are a sign that OLPC is a failure

On the contrary, the idea of leveraging interest in the developed world to help jump-start the program in the developing world has been on the table since Day One. It was suggested by community members as early as January 2006. We hesitated to launch such a program until we authorized mass production. Now that we are about to ship laptops, it is an appropriate time to reach out. These programs will enable a broader community to participate in our mission to give an opportunity for learning to every child. They complement and enhance our government programs.

Peru gave short shrift to planning, implementation and project sustainability

False. A recent Gartner report suggesting that Peru "gave short shrift to planning, implementation and project sustainability"—despite being presented with ample evidence to the contrary. The decision to deploy one laptop per child in Peru was over one year in the making and an implementation plan was begun to be developed from the start. It was not hardware that drove the decision, but sound pedagogical backing. Peru has more than 20 years experience in Constructionist projects implementation. The Minister himself holds a Master’s Degree in Educational Computing and Technology by the University of Hartford, the same university that provided the training for the original implementation team of the Omar Dengo Foundation in Costa Rica. One laptop per child is not seen as a technology project but a pedagogical one; laptops will be distributed as “educational material” in the same category as books and notebooks; the Ministry has ample experience in the logistical aspects involved in attending the needs of 46,000 schools around the country and has the budget for distribution and support already included in its yearly operational budget, in fact, distributing the laptops will probably be less expensive than their current budget. The congress legal approval was required because of the nature of the project and just for what is new. Logistics, support, and distribution is part of their day to day responsibility and will not require any additional law to be supplied.

The Gartner report imputes that "financial provisions are needed in the areas of distribution, transportation, storage, infrastructure, implementation, teacher and student training, content development and permanent technical support", implying that these provisions have not been considered by the Peruvian government. This too is false. Of course these provisions are of critical concern, and consquently, distribution, transportation, storage are included in the budget; the Ministry staff has the proven capacity to manage a project as large as—even larger than—this one successfully. Regarding infrastructure, the project will take advantage of the synergy between Education and Telecommunications sectors, Education is already providing Internet access to more than 3,000 schools, one thousand of them through VSat in remote places. The Telecommunications sector has a plan for rural Internet access in about 3,000 rural communities; a separate budget will not be needed for these since it is part of another sector with whom Education has worked closely.

It is irresponsible of Gartner to suggest that Peru has been remiss of careful planning. Teacher and student training are the of key importance: the pilot projects have shown almost immediate appropriation of the OLPC pedagogical concepts by rural teachers, who were able to seamlessly integrate the laptops into their teaching styles; the Pedagogical IT unit at the Ministry has prepared a training unit to be included in the deployment plan; the experience has shown children barely need any training above what the machine itself provides, the training unit for teachers includes a “first steps” section for children. On top of that, online tools have gotten so good that we feel comfortable teachers will be able to exploit those resources as well. The Ministry portal offers lots of advice already available. Permanent technical support was implemented and has been in place for several years now. A shift of focus is all that is required and that is already under way. Peru also has a provision for spare machines to cover DOA shipments and spare parts. Each shipment will include extra machines, included at no additional cost.

Finally, Gartner presumes to make recommendations as if these topics had not been long considered:

• Plan thoroughly before acquiring systems.

Who would disagree with this recommendation? That is precisely what Peru is doing.

• Guarantee resources to effect implementation and sustainability.

Same comment

• Ensure that technologies fit the educational model, not the other way around.

That is exactly what was done in Peru.

• Make sure that selected providers and technologies are stable and R&D will continue for the success of future phases of the projects.

There are contracts in place to guarantee this. Further, Peru and OLPC are involving local universities to ensure local development and ensuring that Peru is part of a growing international development community.

• Execute user profiling, as this is key for the deployment and successful utilization of the systems.

Peru has in place a detailed profile of children and teachers and has planned the pedagogy and technology to fit their identified needs. Peru also has an evaluation and monitoring strategy.

OLPC is against competition

False. When Intel Corp. “inadvertently” leaked to the news of its departure from the One Laptop per Child (OLPC) Association, the company branded OLPC as anti-competition. Intel creates a false impression by suggesting that the essence of OLPC’s anti-competitiveness is its objection to the introduction of competitive products. In fact, OLPC’s objection is to anti-competitive practices, not products. To object to unfair competitive practices is hardly a stance against competition.

Having "Intel inside" makes a difference

False. We are often asked to compare the Intel Classmate with the XO Laptop. While I've never met a child who cared whether or not she/he had "Intel inside", there are some differences:

• The Classmate screen is woefully inadequate for reading.
• The Classmate has a shorter expected lifetime.
• The total cost of ownership for the Classmate, even discounting that the purchase price is greater, is much higher.
• The Classmate doesn't (as of yet) support mesh networking for peer-to-peer collaboration in the classroom.
• Does the Classmate have a webcam these days? Initially it did not.
• Is the Classmate is a laptop? Has it been certified for on-the-lap use?
• The Classmate display contains mercury.
• The Classmate battery burns at 1000C. The OLPC XO laptop battery burns at 100C.
• The Classmate battery has only 25% of the lifetime of the XO Laptop (500 cycles vs. 2000 cycles). To replace the battery three times over the lifetime of the machine would be almost the equivalent cost of buying a second XO Laptop.
• The Classmate cannot be used in the sunlight.
• The Classmate does not pass a 5-foot drop test.
• The Classmate is not spill-proof.
• The Classmate doesn't have support for direct sensor input (See Measure#Sensor Input into Turtle Art for an example of what you can do with the XO laptop).

The Classmate does have a Windows XP port. Microsoft has announced that they are working on a Windows XP port for the OLPC, but they have not yet confirmed whether or not they intent to commercialize it. But the port is quite stunning. The larger display makes a huge difference in the experience.

A detailed comparison between the Classmate and XO Laptop can be found here.