No Comparison: OLPC XO is the Low-Cost Laptop for Teaching

   
   
   
   
   
olpc enter key
Enter the OLPC XO debate

I am Winter, and I am sorry to say, but I think that most of the criticism directed against One Laptop Per Child here is misdirected. I have read about Nicholas Negroponte's, or OLPC's, arrogance. But Steve Jobs, Larry Ellison, and Bill Gates are all arrogant, and their products range from brilliant to pathetic. Arrogance is irrelevant to consumers.

I read about the "error" of fixating on $100 price tag, but that is the price point for primary schools. A more expensive laptop simply will NOT reach 6-12 year olds. The latest fall in the dollar can make up for some of the price increase, but I do think the OLPC is right that a higher price will simply make the laptops unavailable to younger children.

I read demands for experiments, but large scale experiments on children are simply out of the question. They have not been done in the past (on any topic), and won't be done in the near future. And it is fun to discuss the technical challenges of mesh networks. But as long as we lack contributions from people who actually have any knowledge about the technical sides, this remains just speculation.

There seem now to be more options for computers in the classroom. We do see a difference. The OLPC is also targeted to young children in regions where there is a severe shortage of teachers. The other projects are targeted to high school children that have adequate teachers and facilities. This difference IS important to the number of children that can be reached.

So, I do think that there are better subjects for discussion than Negroponte's interpersonal skills and speculations about mesh technology not yet implemented. I have written about the importance of gains in productivity in teaching earlier, so I will limit myself here to the education process itself.

olpc kiddies
Important OLPC stakeholders

What do I think is important? The fundamentals of education:

  1. Safety (self evident)
  2. Interaction (that is what a classroom is for)
  3. Activities (Children are bad at passive learning)
  4. Knowledge (what is taught)
On each of these point there are important decisions to make that can sink any project if you get them wrong. I am interested how the OLPC does in these matters, and how it stacks up to the other projects. We DO want that the children get the best of breed, don't we?

Safety
In the context of the laptop projects, safety has two aspects, physical and software security. The physical security tries to use safe (non-poisonous) materials. That has been central in the design of the XO, much less so in other laptop projects. Both the XO and Classmate have anti-theft measures build in. The OLPC has set up a comprehensive infrastructure for this. I have no idea how well the other projects do in this area.

Software security can be judged on what is protected. In MS Windows, the security tries to protect Microsoft (eg, wrt liability) and the user is on her own. In Unix-like systems, the computer is protected, or better, the owner can protect the computer. The OLPC model, Bitfrost, protects the user. For consumer systems, Bitfrost is comparable to the models used in, eg, aircraft maintenance: You should only be able to do something wrong if you really know what you are doing. All the other projects just use what is in the software they install, eg, XP.

Interaction
Interaction is communication, conversation, and collaboration. Good teaching always is an extensive two-way interaction. Between teacher and student and between students.
The Sugar UI tries to be an integrated interaction platform. This extends the conventional email/Instant Messaging/VoIP model. Sugar is revolutionary in that it integrates peer group collaboration (and communication) on all levels.

For instance, sketching together is build into Sugar. MS Windows on the Classmate seems to don't go much further than a client-server model using SMB shares. For many tasks, special programs can imitate single aspects of Sugar. But for XP and Linux, there is no real integrated model.

olpc activities
Multiple OLPC activities

Activities
No, using Office is not an activity. Making music, drawings, projects, and playing games and hide-and-seek are activities. Activities must make sense to the children to be useful. Teaching is making children doing activities that help them learn.

The classical model of computing is processing memory-less, flat and unstructured, information. The root example of this is are the linear text and the relational database. Sugar is built completely on structured activities and portfolios. For instance, in Sugar, all activities are collaborative by default and have a history. (on a personal note, MS Office is unfit for structured projects, as anyone who tries to write a book in it discovers)

The classical model is build around the file+application concept, where a file is an unstructured linear array of information interpreted by a single application. Attempts to let more than one application work on the same file show the "one file - one application" history so much that most users only combine the output of different applications after they finished everything.

Modern efforts to free information from the one-application disadvantages are, eg, the Open Document Format. Noteworthy, Microsoft is doing everything to prevent consumers from using ODF, and pushes its own, application bounded, model, ECMA OOXML.

I think Sugar is way beyond any other project with respect to integrating activities with interaction. All the other projects are still based on the old "one file - one application" paradigm.

Knowledge
Central to knowledge is the realization that it only comes from conversations (actually, discourse). Knowledge is more than information. In education, this translates into instruction by teachers, and books that explain more than inform.

In this respect, there is little difference between the laptop options. Whatever educational material you can put on one, you can put on the other. The only edge in is how good the laptops are for reading books and using educational materials. In this respect, the XO's large, direct light readable screen makes it superior.

olpc photos
Center of attention

So I think there is a lot of importance that can be discussed. It still surprises me that it is the cheapest option, the OLPC XO, that has the edge on all four important aspects in my list. In most cases, bad design has lower price as an excuse. Here I can only see haste and lack of interest as excuses.

One Laptop Per Child has its distribution model as a huge disadvantage. However, both the production and use model (mesh network) of the XO depend on large batches being made available. Individual XO's are much less useful due to the network effects.

It is these network effects that I think Intel tries to disrupt with the Classmate. And I suspect that Intel (and MS) will refuse to make the Classmate and XO integrate in any meaningful manner for precisely this reason.

Related Entries

33 Comments

I think you are absolutely right, I am all in favor of the OLPC (the project, the software and the hardware).

The problem I see is that one country has to do the big step of ordering the OLPCs in a quantity of 250,000. Once this has happenend, I am sure others will follow. But at the moment everybody is holding back for different reasons. MOUs are a good thing, but in the end, contracts have to be signed and money has to be paid.

And that doesn't seem to happen soon as more competitors come into the field and there is more choice for the countries. I don't think the September timeline for the start of mass-production is going to be a reality.

Competition is good. But in this case, I still have doubts whether Intel is in it for real.

All realistic estimates about the production costs of their machines are in the $350-$400 range, but all quoted prices are around $200-$250. This discrepancy has never been explained by Intel, and they never really denied the public claim of Negroponte that they heavily subsidize their laptops. And they HAVE publically allocated $1B for pushing their solution. These heavy subsidies would also explain their unwilligness to take large orders.

So, my question is, will Intel deliver laptops for $200 even after the OLPC fails? It is in NO way in their commercial interest to do so.

Winter

Btw, exporting goods below cost is called dumping and is illegal under WTO rules.

So, if true, this would not be a case of, Welcome to the NFL, but just plainly illegal.

And I still have not seen a single clear denial from Intel.

http://www.redorbit.com/news/technology/957934/laptop_wars_it_could_transform_life_for_billions_of_children/index.html

Winter

Rob Winter,

although I quite agree with most arguments in your 4 points I do not agree with the introduction and the conclusion deducted from the 4 points.

Your 4 points are about which laptop is the fittest for education. This is obviously still the XO. But even more important are the following two questions:
a) Why do the target countries hesitate to order XOs eventhough it is obviously the fittest laptop for education?
b) What could OLPC do about it?

"...I think that most of the criticism directed against One Laptop Per Child here is misdirected. I have read about Nicholas Negroponte's, or OLPC's, arrogance. But Steve Jobs, Larry Ellison, and Bill Gates are all arrogant, and their products range from brilliant to pathetic. Arrogance is irrelevant to consumers."

Prof. Negroponte's arrogance during the last year (I don't know whether he has stopped doing that in the meantime) was caused by his lack of sympathy for the fears and responsibilities of the decision makers whom he even ridiculed. And that is definitely most relevant.

"I read demands for experiments, but large scale experiments on children are simply out of the question. They have not been done in the past (on any topic), and won't be done in the near future."

If no pilot projects yielding solid conclusions can be done then the whole roll out will become itself one huge experiment risking hundreds of millions that are not in possession of OLPC but of the countries' tax payers. This would simply be irresponsible and this is in my opinion the main reason why the countries have not ordered yet. So I keep demanding scientific pilot projects that yield solid conclusions.

"It is these network effects that I think Intel tries to disrupt with the Classmate."

Yes, it looks indeed like Intel tries to do that. But why are they successful with that? Because their price is still higher than that of an XO? Or because it is obvious to anyone that the XO is superior in education? No, Intel is successful with interrupting OLPC because of OLPC's large scale project risks that OLPC has so far neglected or even refused to work on. Any alternative looks less risky than OLPC because they are ready to start small.

If OLPC wants to find a way out of the current crisis they have to change their attitude towards risking somebody else's money. It might even be that OLPC is so certain about their success that they don't even understand the sorrow of the target countries.

First OLPC has to learn to become sympathetic with the countries responsibilities and fears and only after that the educational superiority of the XO is becoming decisive.

Sometimes I have wondered why the entire world is blind to the technological and educational marvels in the XO. And, every time it happens, the same thoughts come to mind:

I can't remember the last time anyone ordered 1,000,000 units of any product based on a not-fully-functional prototype:

1. Mesh Network Field Testing TBA.
2. Power Generation Solution TBD.
3. Mass Storage Solution TBD.
4. Final Energy Consumption Specs TBD.
5. Classroom Implementation TBD.
6. Total Cost of Ownership TBD.
7. Price Per Unit TBD.
8. Tech Support TBD.
9. School Server Availability TBD.

Then, there is the very normal fear of doing business with start-ups: what happens to the tech support, return policies and the laptop itself if OLPC folds after 1 or 2 years, for whatever the reason? It is the same fear that drives people to buy expensive products from mainstream merchants, instead of buying them from the little corner business that may or may not be there next month.


Mickey R. Duff (aka as Martin's)

"I read demands for experiments, but large scale experiments on children are simply out of the question. They have not been done in the past (on any topic), and won't be done in the near future."

You are partially correct in saying that such experiments were not done in the past. However that is not an excuse to do them now. XO can make wonders, I sincerely hope so, but it is still untested. I'll give you an example.

A new breakthrough medicine is developed, and from the tech specs it will cure the worst possible disease. Everybody is raving about it. However there is a catch: you need to be ordering an enormous amount, up front. There is no time/need for testing, in fact on paper the manufacturer is so convinced that it will work to the point of dismissing completely testing. Sure, there are technical details that makes it "probably" more useful than other more traditional medicines.

How do you feel about such a product? Don't you think that maybe a scientific test of it would be appropriate? The FDA requires it, and for a reason.

I said in the past and I will said it again. When we are talking about education there is not really a recipe for success. Or if there is, than we would have solve the literacy problems we have. I know I am pushing this to the extreme, but not so much. When the US federal government sponsors any educational project, testing is essential and required, because it's simply tax payer money. There are very clear guidelines about this.

So why then the XO should be exempted? To rephrase one of Winter statements:

"But as long as we lack contributions from people in the field who actually have any knowledge about the educational sides in developing countries, [the effectiveness] remains just speculation."

I think the comments from Rob Winter about Intel offering its competitor solution to the OLPC at below cost can be put to rest at this point. ASUS announced the pricing for its new ultra portable laptop. Specs are similar to the Classmate / OLPC or better. Release date in August of this year price is $199. Seems like it is possible to reach the $200 when the display gets scaled back in size. See here: http://gizmodo.com/gadgets/pocket-browser/asus-eee-pc-gets-pricing-and-release-date-266846.php

As well arguing that a fall in the US Dollar against other currencies makes up for the higher price is a non argument in my few. The cost for labor and materials on the international markets are adjusting to the devaluation of the US dollar and therefore increased the dollar price and keeping the equivalent Euro and Pound prices stable.

The question is what are you comparing it to? Your disease model has an interesting catch -- suppose people already have "the worst possible disease", and there's no treatment yet. Your choices are to try an incompletely-tested new medicine or suffer from the disease -- now which one do you pick?

It's not like OLPC is competing against something that's already "good enough". In this case, we're talking about fundamentally broken educational systems where the OLPC might genuinely help. It's really hard to imagine how this cure could be worse than the disease -- and it might well help a lot.

Bravo to Winter for countering some of the incessant negativity here.

Kent - "Your choices are to try an incompletely-tested new medicine or suffer from the disease -- now which one do you pick?"

Well, to be blunt, it depends on how much faith you have in the "new medicine." It's been my experience that quick fix miracle cures work about as well in development as they do in getting rich or thin. In fact, incompletely tested "cures" can end up worse than the disease. What if, despite the best efforts by OLPC, these turn into hot black-market commodity items, putting children at risk? (Feel free to imagine your own horrible remotely-possible scenarios). The real risk is that the simply fail to revolutionize education in the promised ways, causing governments to be even more averse to trying new technological solutions.

Development is slow, hard work. Trying to make it into an emergency "people are dying we've gotta fix this now!" situation won't, in the end, help anyone. A few months worth of in-country pilot projects has the potential to catch so many glitches and prevent so many currently unforeseen difficulties. The only reason not to do it is for marketing/commercial reasons - the need to make the 3Mil order or fear that the pilot projects won't perform to the promised levels.

I am not being negative, just trying to be pragmatic.

I am all over trying new products. That is in fact what testing is all about. My point is that very often untested medicines turns out to be placebos. Sure they don't hurt, but they don't cure it either. So you think is good to have poor countries to pay for something that may not cure their problems? All points Winter has correctly listed indicate that the product has a lot of potential. But as far as we all know, this does not mean that this potential can effectively translate in improving education in developing AND in developed countries.

Again, I have the highest expectations for such project, and I seriously hope they will be successful. As a scientist, I was trained to think critically to anything I am exposed to. The scientific method tell us, once you have a theory, you need hard proofs to confirm it. If the OLPC can prove itself to be really making a breakthrough in education it will have all my support and enthusiasm.

"I read demands for experiments, but large scale experiments on children are simply out of the question."

It doesn't have to be in "large scale", Rob.

A few pilot projects - say, 10 of them - in different countries with 1,000 units per project over 6 months would be perfectly fine to settle most of the issues that Mickey listed:

1. Mesh Network Field Testing TBA.
2. Power Generation Solution TBD.
3. Mass Storage Solution TBD.
4. Final Energy Consumption Specs TBD.
5. Classroom Implementation TBD.
6. Total Cost of Ownership TBD.
7. Price Per Unit TBD.
8. Tech Support TBD.
9. School Server Availability TBD.

I think your error is in not dealing with those issues in a scientific way, as Nick points out.

You are confusing OLPC goals with achievements:

1. Goal: "We want a working mesh network"
Achievement: not there yet.

2. Goal: "We want a human-operated power generator"
Achievement: not there yet.

3. Goal: "We are planning to implement a Mass Storage Solution that includes Google"
Achievement: not there yet.

etc., etc.

This is the obvious reason for countries not being willing to place any orders now or in the foreseeable future. Countries, quite understandably, would love to buy a finished and tested product, specially one with so many new features.

Now, don't get me wrong; I don't know how good or bad the XO is; I don't know if the XO will ever live up to its promise or even surpass it.

My point is that with Windows, for example, you know exactly what you are getting, for better or worse; we have tried it, some love it, others despise it; that's ok.

With the XO, we don't even know how it works or if it works; it is a completely untested product. In the end, the XO might be 100 times better than anything else in the market; it could be 100 times worse; we won't know until it is put to the test under real-world conditions (a couple of un-documented trials with 30 kids don't count, of course).

I think countries a very wise in withholding puchasing decisions until they can compare different products, regardless of creator.

Bob,

"With the XO, we don't even know how it works"

OLPC project, unlike other ones we discussed here, is fully open - there's plenty of info if you really interested.


"it is a completely untested product"

XO's OS is Fedora based, ie Linux.

GUI library - GTK is fully portable (quite a few programs, like Gimp, can run on Windows thanks to using GTK) and the main base library for Gnome (as used in Red Hat, Fedora, Ubuntu, SUSE and many other Linux distributions).

Main programming language - Python and increasing popular (eg. most of new software for Ubuntu is written in Python).

The basic activities XO provides is:

1) Writing - using Abi word processor,
2) Painting - using Tux Paint program,
3) Internet browsing - using Firefox-based browser,
4) RSS reader - using PenguinTV,
5) Music composing - using TamTam,
6) Multimedia programming - using EToys with Scratch being currently ported,
7) Games - any game (or other apps for that matter) using GTK should be easily ported.

None of the above is in any way 'experimental' as it's based on well established software...What is new, and should be treated as a *bonus* compared to other contenders, is all these is in a hardware-software package specifically targeted at small kids in developing countries (and eventually others living in similar conditions) who need to use a laptop outside the classroom (hence a special attention paid to case, battery and screen).

And of course:

0) Reading ! - with its eReader program, folding down dual mode hi-res screen, power saving 'eReader' mode, other contenders don't come even close....

You guys keep confusing software (Sugar) with the hardware (Classmate/XO). There isn't any particular reason you couldn't port Sugar on the Classmate or ASUS laptop. It's certainly not hard to run in a virtual machine on any PC. The real wins for the OLPC are the dual mode screen, the mesh networking, the power consumption, and the durability/made for children aspect. It's a much nicer piece of hardware than the Classmate prototypes that have been shown so far.

Delphi,

I guess we have different definition of "tested". Sugar is well tested if you mean stability of the programs and applications. But this is a geek/developer perspective.

Sugar is untested in terms of their effectiveness in education. In other words:

1) Writing - using Abi word processor
How is this is going to improve literacy? How is this "better" than say pen and paper or other word processors?

2) Painting - using Tux Paint program,
See above

3) Internet browsing - using Firefox-based browser,
Too generic. Any computer can have Firefox on it. What is special about sugar? Or better again, how is sugar making the online experience better suited for education?

4) RSS reader - using PenguinTV
See above

5) Music composing - using TamTam,
See above.

6) Multimedia programming - using EToys with Scratch being currently ported,
Great apps. How do you integrate them with an educational curriculum? For example, how do you integrate these apps with teaching physical sciences?

7) Games - any game (or other apps for that matter) using GTK should be easily ported.
Please list good educational games that can effectively be used as educational tools.

The two thing that sugar provides over a conventional laptop are the sharing and the new file system based on a journal. Those factors are what makes the XO special and """potentially""" great. Until we will know that using those features in combination with the apps you mentioned above, will indeed (as we all hope) very effectively improve education, I will consider those features untested in an education context.

Nick,

I think you misunderstood my post - I didn't claim that the software available on OLPC's XO is 'better' than what you can run on Classmate PC or Eee PC. The reasons I listed them (in my next post, which for some reason didn't show up yet, I added, perhaps most important one, eReader) was to show that the software XO is going to use has been well tested and is well established. The main (apart from additional and very important, as you say, collaboration features you've mentioned) is that with its dual-usage high resolution screen, they way it's built (notice the big vents on Classmate PC to accommodate fan...yuck!) and its low power usage (especially in eReader mode) is so much more suitable for string-pullers (as Martin aptly calls them) than anything else - I really can't understand why it's not obvious for everyone! And they still have managed to have the price cheaper than the other contenders. As Winter says, there's simply no comparision...


As to software in education:

"For example, how do you integrate these apps with teaching physical sciences?"

Both EToys ( http://wiki.laptop.org/go/Sugar_EToys) and Scratch ( http://scratch.mit.edu/ ) are ideal for all kinds of simulations ( I remember seeing electronic circuit simulation).


"Please list good educational games that can effectively be used as educational tools."

You can find some here: List of educational software (including educational games)
( http://en.wikipedia.org/wiki/List_of_educational_software )

but my favorite, and the most suitable, I think, for small kids is GCompris (shouldn't be a problem running it on XO)
(http://gcompris.net/-en- )

Time zones and other obligations prevented me from participating up to now.

I see four main threads:
1 The XO is untested. We should test it first.
That would be nice. However, even a set of deployments of 1000 laptops will only be convincing if there is a control group (see the medical cure developments). Please, calculate for yourself how long this must take (6 months at least), how many people should be involved (3000 pupils, 200 assistants, 50 specialists), and how much time it will cost to get the plans through the legal and ethical commissions (1 year minimum, I guess 2 years). You are experimenting on children, remember. And everything can only start after the finished product has been delivered. At which time it cannot be changed anymore.

"When the US federal government sponsors any educational project, testing is essential and required, because it's simply tax payer money."
The introduction of computers was not tested in the USA. There was just an enourmous amount of tax money used to order them. Just as the studies on pre-school education where happily ignored. And don't start on USA curricula, with many subjects being severely limited and distorted by religious bias.

2 The hard and software are not ready, not tried and not tested.

About the stability of HW and SW. Anyone betting Sugar will be less stable than XP? Anyone betting the XO will be more fragile than the Classmate, which doesn't even have a sealed keyboard?

The discussion about the Sugar interface reminds me about the introduction of the Macintosh in 1984 (I am that old). Look back to the reviews then. "The Macintosh is not a real computer but a toy", "MS Dos is the future". Real (paraphrased) quotes. Alan Kay was around then (at Apple), he is doing it again it seems.
http://www.windley.com/archives/2006/02/alan_kay_the_10.shtml

The software is written/designed by people who have a lot of experience with really large and difficult projects. If they can't complete it successfully, I wouldn't know who could. And NONE of the things the OLPC attempts is actually a vexing technical problem.

In my own projects I hate people who want to have everything ready, tested, and proven half way the project. This is the famous "first step fallacy". So I will not require that from the OLPC. A release date is there because you plan to have things ready at that time, not several months before.

3 What is the value of computers in education?
Please visit a high school in your neighbourhood. All schools in all developed countries use computers in education. In my country, you cannot finish a high school that prepares for higher education without a computer (and probably neither the others). Telling develloping countries their children don't need what is required from rich children is condescend.

4 Why do they not order?
I said in the post that the distribution model of the OLPC is a liability. And I do not see how this can be changed easily.

Without the network effect, the educational value of the XO is much less (especially as Microsoft will actively break any interoperability). Mass production is necessary to get the price down, and to limit the costs of distribution (see the link above).

The team members are travelling the world doing demonstrations. You can find videos of test deployments on YouTube. But I agree, $200M is a lot of money, and the logistics are daunting (and the OLPC project CANNOT organize the distribution).

Some other matters:

"My point is that with Windows, for example, you know exactly what you are getting, for better or worse; we have tried it, some love it, others despise it; that's ok."

You really want to let millions of young children struggle with Windows? With virusses, identity theft, zombie botnets, spam, firewalls? Personally, I would consider that a crime.

And we KNOW what we get with Linux. Linux has been around for 17 years now. It runs on the fastest, biggest, smallest, and cheapest computers. It has been tested and proven beyond anything XP has ever gone through. Every single aspect of it has been studied and discussed to death. And you say only XP, with all its easter eggs and hidden interfaces, its WGA, is known?

The dollar price:
As the dollar falls, the XO, which is produced outside of the USA will become more expensive in dollars. But the clients are also outside of the USA, so they will not pay in dollars anyway. The price should be compared to something relevant to the client countries, eg, the price of oil or coffee, or euro/yen/yuan exchange rates.

Negropontes interpersonal skills
This is not MS, where Bill Gates can ruin every project he likes, the way he likes. The OLPC is run by more people than Negroponte. Personally, I don't care for Prof Negroponte. I care for Kay, Tossati, Greg KH, the people at RedHat and Google, the people designing the HW. They are the ones that make the important decisions. By fixating on the PR face, you are looking in the wrong direction.

"Yes, it looks indeed like Intel tries to do that. But why are they successful with that?"
Maybe a $1B slush fund and decades of experience in illegal competition practices (see the AMD vs Intel anti-trust suits)? We do have a letter from Intel to the Nigerian governement that was using Intel's clout to quite simply lie on everything the OLPC tried to achieve. And I am still wondering what legal uses Intel have for these $1B for $200 laptops?
(about the price, the Classmate comes in several configurations and $200 is the LOWEST list price for which you currently CANNOT buy a Classmate. In this case, seeing them for sale is believing)

And finally, if you look at the title, a lot of faults and shortcommings of the XO and OLPC were pointed out. Which of those "problems" do NOT affect the other options?

Winter

Rob Winter,

"Please, calculate for yourself how long this must take (6 months at least), how many people should be involved (3000 pupils, 200 assistants, 50 specialists), and how much time it will cost to get the plans through the legal and ethical commissions (1 year minimum, I guess 2 years). You are experimenting on children, remember. And everything can only start after the finished product has been delivered. At which time it cannot be changed anymore."

"The introduction of computers was not tested in the USA. There was just an enourmous amount of tax money used to order them. Just as the studies on pre-school education where happily ignored."

I do not quite understand what you want to say in above quotes. Do you think it is too laborious and lengthy to do serious testing? Are you ready to accept gambling several hundred millions of foreign tax money?

If it is so then maybe you should earn your money e.g. in Peru for a while and pay taxes from it there. This would teach you the value of hard earned tax money.

According to your post and comment OLPC has done everything right. All the problems are caused by external "bad"-guys. So what in your opinion should OLPC do now to save their mission? Sue Intel? Or shut their eyes and ears and storm forward and in the end blame again everybody else when it turns out that all the money and effort has been wasted?

Please, tell us: What do you think OLPC should do to achieve their mission?

Roland wrote:

"According to your post and comment OLPC has done everything right. All the problems are caused by external "bad"-guys."

Did I say that? That would be most unfortunate, as I don't think anything like that.

My point in the original post was, that considering the alternatives, the OLPC project has worked miracles. The XO is a technical breakthrough, potentially of the same size as the Apple Macintosh in 1984. Just as Apple messed up their lead with bad implementation (eg, refusing to produce a Mac with 2MB memory), the OLPC can mess up the XO.

What the XO has NOW, at this moment, is better than what Intel can deliver for twice the price. I haven't seen any poster denying that. Just things like "it is not proven". But XP is proven defective and unsafe, and that seems to be no problem for some of the same posters. Not that that is a problem for Intel.

Actually, I see only a single problem that might be traceable to commercial competitors with a criminal track reccord in illegal bussiness dealings (see below).

"Do you think it is too laborious and lengthy to do serious testing? Are you ready to accept gambling several hundred millions of foreign tax money?"

Someone claimed educational theories get tested in the USA. I don't see that happening, so I gave some counterexamples. This is just to show that rigorous, or extensive, test phases in education are next to impossible, and have never been done before.

Moreover, you might want to consider what it means to defend a research experiment involving 1000's of children in front of several ethical committees (one per target and origin country). For those who have not experienced that, my advice would be to keep out of it.

I just want to say, too, that a test involving 2000 children (test and control) might cost you millions of dollars too and several years (I would guess 3 years). And then you will find out you made an error that invalidates the experiments for all the people that don't like you or the outcome anyway. (this is what happened to all those studies on pre-school education in the USA, if I remember well)

"Please, tell us: What do you think OLPC should do to achieve their mission?"

The single failure of the OLPC project is ensuring enough orders, it seems. This is probably to a large part caused by the enormous sums of money involved (5 million laptops ~ $800M). Which IS a lot of money.

What is a concern, is that at least one of the comercial competitors is engaged in guerrilla warfare using large amounts of money, strategical price dumping, and, what could be, fake offerings. Nothing new here.

The result might be that several governements might make a misinformed choice (in both directions). Not because the OLPC is always right, they aren't, or the XO is always the best choice, it isn't, but because Intel, MS, et al. always lie and have a lot of money to make their lies stick.

In the long term, this probably doesn't mean a lot. But in the long term, we are all dead and there might not be an XO around to help out. And the children have to wait years for a new opportunity for their children.

A way out?

That depends on how much orders the OLPC did get, the number of orders needed to start the project, additional funding the OLPC can get, and several other factors. And, as many posters have remarked, there seems to be some doubt about how these laptops can be intergrated in the existing system. On the other hand, I have seen extremely bad IT decisions made by ignorant CEO's and politicians in rich countries. Why should the developing world be better in this respect?

There certainly is a comercial market for the XO, but distribution and marketting would make everything extremely expensive (in general, retail tends to double the price). Funding by "adopt a laptop" might work too. Foster Parents works something like that. At least, the pictures and videos of the children with the laptops would be terrific marketting material.

But I am sure the OLPC has people on board who know much much more about financing of this project than I will ever do. So I will do just as the rest of us, I wait and see.

Winter

Thanks for your answer, Rob.
When I understand you correctly you indeed think OLPC should roll out without additional testing. They should overcome their current lack of orders and the consequential rise of unit price by some kind of additional funding.
(Why not also risk somebody else's money if some target countries don't want to have theirs risked ?? Is it more moral to risk first world charitable funds instead of third world tax money ?? Or doesn't it matter if only somebody is courageous (or stupid) enough?)

I agree it must be rather difficult, expensive, laborious and time consuming to carry out scientifically valid testing.

But I assure you it was a million times more difficult, laborious and time consuming for the citizens of the target countries to earn that kind of tax money that you seem to be ready to spend on so weak justification. That such irresponsibility has been committed before by other countries does not justify to repeat it consciously. This would constitute misappropriation.

By the way it is not necessary and also not possible to silence all criticism by tests however scientific. Of course the loosing competitors will always try to question the results. The natural solution would be if you let the decision makers and their experts define (and maybe also carry out) the kind of tests they accept to justify spending that kind of tax money.

There remains large room for speculation:
We don't know whether they already tried to do that. We also don't know what support OLPC offered for such testing. Perhaps there is already a reduced number of orders just for this purpose. A few thousand per country maybe? Is OLPC willing and able to deliver such a reduced number of units? If not, why not? Might this also be a reason why Intel is offering their Classmate just in the range of ten thousands? Perhaps until more test results are available? We don't know.

Rob Winter wrote the following:

"What the XO has NOW, at this moment, is better than what Intel can deliver for twice the price."

I doesn't mean anything. Stalin was better than Hitler.

"Please visit a high school in your neighbourhood. All schools in all developed countries use computers in education."

This is not a high school project. And, to be quite honest, nowhere in the world are elementary school children using a laptop per student for educational purposes.

"The introduction of computers was not tested in the USA. There was just an enourmous amount of tax money used to order them. Just as the studies on pre-school education where happily ignored."

Very true. That's why third-world countries need to be careful with their money and wait until the OLPC conducts the necessary testing.

"Moreover, you might want to consider what it means to defend a research experiment involving 1000's of children in front of several ethical committees (one per target and origin country)."

Serious projects require a lot of work and dedication. It is easy to be irresponsible, close the eyes and hope for the best. True scientists adhere to scientific methods.

"You are experimenting on children, remember. And everything can only start after the finished product has been delivered. At which time it cannot be changed anymore."

I had to read the above statement twice. Only a madman would advocate such suicidal, irresponsible strategies.

It seems to me that Rob winter's approach is based on a few demented premises:

1. The XO is untested, but that should not prevent countries from spending their money

2. Should the OLPC Project fail, Microsoft, Intel and everyone, except the OLPC team, is to blame.

3. The reason countries are not ordering is that the "bad guys" are forcing their hand.

4. Any person who questions the OLPC methods is clearly in favor of Microsoft, Intel, Asus, and the rest of the "bad guys"

Roland wrote a lot, eg,
"When I understand you correctly you indeed think OLPC should roll out without additional testing. They should overcome their current lack of orders and the consequential rise of unit price by some kind of additional funding."

The point I am trying to make is that scientific tests are expensive, difficult and risky. Yes experimenting on children is inherently risky. And that no educational reform in human history was based on a controlled test on large numbers of children (but I welcome any information that contradicts me on this). And that educational politics are NEVER about scientific test results.

Whatever scientific experiment with test and control groups you will ever perform, it would be a first time in history, by my knowledge, if the results would be really used in setting national educational policies.

What is the alternative?
People, ultimately parents, have to decide what they think is good for their children. It is also THEIR tax money. It was this same motivation that brought computers into Western classrooms. So, what sauce for the goose, is sauce for tha gander.

And I do not even want to suggest that these countries should get the XO's forced or dumped on them. The whole point of the OLPC project is that the countries buy the laptops with their own money. So these are their own laptops and they are in control. This is not our call to decide. If they want tests, they can do them. They certainly have the scientists for it (never trust statistics you didn't compile yourself).

What many of the posters here tell us is they want extra time for the OLPC to do all kinds of test rides and other development. That costs money, lots of it. And producing thousands of "test laptops" for all these educational and field tests cost millions of dollars extra. If you really want testing, you can't claim the OLPC shouldn't have extra funding.

Furthermore, if you really say we can stay the release for more than a year for additional testing and development, you cannot complain if the OLPC then starts mass-producing for other markets, even with subsidies. Because, another year waiting without production means the end of the XO. (see logistics of large scale production facilities). Note that Moore's law will ensure that any laptop used in the tests will not be for sale after it.

So what I want is that the receiving countries make their own decisions on what they think is good for them. And obviously, they will try to do that. And the OLPC is trying to sell the XO to them for good reasons. None of them needs my help for that.

If all goes well, these countries should just ignore all my comments (and I am sorry to say, yours too) and just listen to their own experts. These experts are certainly infinitely more qualified to help making those decisions than most (any) of us. These are not OUR kids, and this is not OUR money.

We are just bystanders watching what happens. And for what it is worth, and for how little I know, I think Brazillian or Mexican children can benefit from the XO. And Nigeria might have the money and incentive to take the risk. I just know too little of the other countries to even make an uneducated guess.

Still, I have not seen any contradiction for Wayan's title on my post, that says that the XO is currently the best of the laptop options. This qualification holds even when you don't agree with me about whether laptops are useful in educating poor kids in developing countries. It even holds if you despise the complete OLPC team.

Winter

"Still, I have not seen any contradiction for Wayan's title on my post, that says that the XO is currently the best of the laptop options. This qualification holds even when you don't agree with me about whether laptops are useful in educating poor kids in developing countries. It even holds if you despise the complete OLPC team."

Being the best of the bunch doesn't mean it is good. As Totico Peluo so beautifully put it to you before, Stalin was better than Hitler, but that didn't mean that Stalin was any good.

We have to wait until Negroponte conducts the necessary tests. Until then, nobody knows how good or bad the XO is.

"These are not OUR kids, and this is not OUR money."

Not true. Some of us are writing from target countries, so, unlike you, we have great interest in determining what's inside Prof. Negroponte's "trojan horse".

My name is Pancho. I love my poncho, mister.

"Furthermore, if you really say we can stay the release for more than a year for additional testing and development, you cannot complain if the OLPC then starts mass-producing for other markets, even with subsidies."

Keep it real, cowboy. Prof. Negroponte has no "other markets".


"what I want is that the receiving countries make their own decisions on what they think is good for them"

That's exactly what they are doing - that's exactly why they are not ready to place any orders. Countries have spoken. Their silence screams: "show me the beef, 'Ponte"

Rob,
this sounds a lot more sensible to me now.
The countries are responsible for their kids and their money. They are in charge. They did not yet place large orders because of the insufficient knowledge base for taking responsible decisions. Not because of "bad guys" as OLPC is publically claiming.

OLPC does not need large extra funding for testing if they let the countries do it and sell them low numbers of test laptops at their actual cost of low volume production. I would not oppose sales of commercial variations of the XO in the meantime until the countries' large scale orders are placed. Probably the current funders of OLPC have a different opinion.

So far it is obvious that the XO has the most educational potential. But the competition does not sleep. There might be improved models in the near future. I do not despise the OLPC team. They brought about the movement in school computing which is very good. And I hope they make no serious mistakes in order to guide this process of change to a successful end. Only now there are more players trying to force the change in different directions. Now even more wisdom and foresight is required.

>>>>>>>>>>
Being the best of the bunch doesn't mean it is good. As Totico Peluo so beautifully put it to you before, Stalin was better than Hitler, but that didn't mean that Stalin was any good.
>>>>>>>>>
"These are not OUR kids, and this is not OUR money."

Not true. Some of us are writing from target countries, so, unlike you, we have great interest in determining what's inside Prof. Negroponte's "trojan horse".

My name is Pancho. I love my poncho, mister.
>>>>>>
"Furthermore, if you really say we can stay the release for more than a year for additional testing and development, you cannot complain if the OLPC then starts mass-producing for other markets, even with subsidies."

Keep it real, cowboy. Prof. Negroponte has no "other markets".
>>>>>>>
"what I want is that the receiving countries make their own decisions on what they think is good for them"

That's exactly what they are doing - that's exactly why they are not ready to place any orders. Countries have spoken. Their silence screams: "show me the beef, 'Ponte"
<<<<<<<<<

We don't really know, do we? Some countries might have signed up.

But I sense some anger in your response towards the OLPC here? Why is that?
The OLPC and XO are in itself not evil projects. The people in the OLPC project did do their best to help out. They did work miracles. Maybe you think it was not enough, but that is no reason to be angry on them?

Winter

"We don't really know, do we? Some countries might have signed up."

Elvis is alive. Spotted at Burger king.

"But I sense some anger in your response towards the OLPC here? Why is that?"

Childhood issues, clearly.



"The OLPC and XO are in itself not evil projects. The people in the OLPC project did do their best to help out."

Even good people need to test they products before anyone will buy them.

"They did work miracles."

And then, just like Christ two thousand years before him, Prophet Nicholas multiplied the computers. His people, however, did not bite.

"Maybe you think it was not enough, but that is no reason to be angry on them?"

Microsoft, Intel, Asus and the rest of the "bad guys" have taken wireless control of my pea-sized brain.

"If all goes well, these countries should just ignore all my comments (and I am sorry to say, yours too) and just listen to their own experts. These experts are certainly infinitely more qualified to help making those decisions than most (any) of us. These are not OUR kids, and this is not OUR money".

This is indeed the most important point of the discussion. We can talk, or not, about testing (either comparative testing of different platforms, or specific testing in a particular environment). The fact of the matter is that if the customers of such product don't have a word about the platform itself, so that it fits their needs, the all product is pointless. We, westerners sometimes have the presumption of knowing what's best for everybody. We produce products so good that they must trust us that they will simply work. If that would be true, the US education system would be the best in the world, which it isn't by far.

Testing is expensive. Having a machine that may NOT suit local needs is expensive. I am sure people will find use for them in any case. But that kinds of defeat the initial purpose.

My question to Winter, whose argumentative writing I simply love. You should be OLPC spokesperson ;-), and I am saying this genuinely and honestly.

Maybe I am not informed, but how many "experts" from the developing countries have been involved in the development of the XO? Was their role central?

Well, it certainly feels good to see my basic arguments taken up by others for debate and (hopefully) improvement. I don't mean to take credit for what should be self-evident, of course.

I'd like to respond to the invocation of Moore's Law by Winter, who says that "any laptop used in the tests will not be for sale after it". While this is true in the sense that the laptop will not be built of the same components, it does not follow that the laptop will necessarily change in important ways.

Moore's "law" is not in fact a law, operating through the processes of nature, but is in fact a metric, implemented by intentional human activity. To see this, simply append the following words to it: "no matter what the cost". As that cost becomes higher, the curve flattens out.

In fact, the same laptop can be made with vastly different technologies. Transistor density is not the main determining factor in the performance of the equipment, just in its cost.

I have only two questions.

1. Does the OLPC $100 computer exist?

2. How can I get some?

I am going to be going on volunteer work to Haiti and I would be willing to buy a few for the children.

Natalie,

Yes, the OLPC XO exists, but you cannot buy one (or a few) for Haiti. The One Laptop Per Child program will only sell multiples of 250,000 to governments.

Natalie,

I'll be more exact than wayan in answering your questions:


1. Does the OLPC $100 computer exist?

The OLPC $100 does not exist. OLPC's curent machine in development has a price tag of $175 and it is not finished or tested yet.

2. How can I get some?

Order at least 250,000 and wait until they are finished. That's the current acquisition policy by OLPC.


Dick Einstein

Natalie might be better to invest in the new laptops coming from Asus. The Eee PC would seem to be a better deal for people only wanting a few low cost laptops for children.

http://www.asus.com.tw/news_show.aspx?id=7317

Except that they're not available yet and Asus is being coy about shipping dates.

Natalie, you might want to research how well conventional laptops do in a tropical environment. I'm going to guess "not very" but people are endlessly inventive and may have come up with solutions to the problems of using a laptop in the tropics.

Close