Ivan Krstić Says Negroponte's Wrong About Sugar and OLPC

Not many days ago, we mentioned ZDNet's interview with Nicholas Negroponte, in which Negroponte had some harsh things to say about Sugar and its connection to the slower-than-hoped uptake of the XO. Ivan Krstic (formerly head of the OLPC's security innovative subsystem) responded to Negroponte's claims, which he says are "nonsense." Among other things, he mentions that Sugar "was the name for the new learning-oriented graphical interface that OLPC was building, but it was also the name for the entire XO operating system, one tiny part of which was Sugar the GUI, and the rest of which was mostly Fedora Linux."

Comments

entirely not the problem

[wizardforce]

Correct. The XO's problem isn't sugar or as far as I can see really anything to do with its specifications but rather how it was sold and marketed.

Re:entirely not the problem

[StreetStealth]

Getting OLPC off the ground was a huge undertaking whose backers did not all share the same vision.

The whole point of Sugar was to make the XO a universal, personal educational computing device, free of the cultural barriers and prejudices that are inherent in something like Windows. The people who pursued this vision of OLPC were the idealists.

Then there were, to use a generous term, the pragmatists. They didn't see the use in building a new, universal education platform. To them, the developing world may well have just been millions of children waiting to grow up to work at offshore call centers, and getting them familiar with The Way The World Works was the first priority.

Obviously, the latter won, and to be honest, I don't think their tamer, "more realistic" vision of OLPC will ever make the same mark on the world that a Sugar-powered XO would have.

Re:entirely not the problem

[Darkness404]

The problem was the unwillingness to tap the first world. Back in 2007 you can bet that the would would have somewhat embraced a laptop that cost only $230, even if the specs were somewhat pathetic. Add that into the fact the machine is nearly indestructible and can be read in sunlight and you would have had a fast selling machine and made $30 on top of it. But of course the OLPC said no to that and instead came up with "give one get one" where for $400 you could effectively get the same laptop and make a $200 contribution to the OLPC fund. Hm, would I rather buy a $400 durable laptop that has about the same specs as my ancient PIII locked in my basement that I never use, or buy a laptop with a decent CPU and a usable amount of RAM.

Just look at how well the iPod touch sold and it cost $300 in the same time period and at launch had 0 apps. An iPod touch with a keyboard would have sold very well when priced even lower. The unwillingness to sell the OLPC computers to the first world was a huge mistake.

Re:entirely not the problem

[Phoe6]

You probably are right. I have heard Negroponte a couple of times and have had the impression. I also have a couple of facts to illustrate that his vision is not shared by prominent people.

1. India did not participate in the OLPC project itself. If it had caught in India, it could have been a different story altogether. There are thousands of kids in India, who in couple of years come to grasp with technologies. Government of India, did not feel that OLPC was a viable thing to do for India schools. Instead of backing off, OLPC should have worked out a solution for Indian market. Considering that India is a third world country and where technology catches up pretty quickly and students catch up with the learning curve easily (British gift of English and Infrastructure being one reason)

1. Even before OLPC initiative, Negropontes Media Lab initiative suffered in India. Then minister Arun Shourie (who is a recognized leader and thoughtful person in his own right), felt that Negropontes Intiative's were not very convincing. Even if it had a tag of MIT.

Also I tend to trust Ivan KrstiÄ more than Negroponte. Those were the guys who sweated it out to bring OLPC. It was sad to see all the tech guys back off the project.

One more thing, If you beat your drums before you have something, you are basically full filling your ego, nothing else. OLPC did that from the first day.
It was also called 100 dollar laptop. Is it available for 100 dollars yet?

Alternative

[Toonol]

I think they should have looked more toward the Nintendo DS as inspiration, honestly. That thing can survive in 3rd world countries no problem; it survives 8 year olds. It's powerful enough to browse, the touch screen allows for a ubiquitous interface; just add a stripped-down linux and allow it to read/run from external memory SD and USB drives.

Wingnut

[Improv]

A far bigger problem than Sugar, if sugar is even a problem, is having a wingnut leading the company. Negroponte's most visible activity wrt the OLPC is to torpedo it. Ditching him would be a much better start than ditching sugar.

Re:Feature creep killed the XO

[Locutus]

IIRC, they didn't do _that_ much for Microsoft and pretty much just added the SD card. But, they did let Microsoft take up their time and even some of their office space. It really wasn't an easy hardware kit to build considering it was supposed to work outside, by kids, and be very energy thrifty. Nobody has even come close to what the XO does and only the Kindle can be read outside in full sunlight.

The Microsoft stuff misdirected the marketing of the project once it was determined that Microsoft and Intel successfully blocked many of the sales they had MOU's for.

LoB

True

[HalAtWork]

Part of OLPC had always been that the entire software stack could be modified, and that users could learn about it and share ideas to make their own platform better.

Re:True

[Belial6]

I think the part you are thinking of is the part where, any criticism of the system by pointing out that it doesn't do a job well is countered by "that isn't what the system was designed for". I can't count the number of times that the failings of the OLPC as a 'learn to program' platform was countered with claims that it was never intended to be a tool to teach programming. In fact, the OLPC website never really said what they were really trying to accomplish. Just a bunch of marketing buzzwords.

Re:True

[westlake]

Part of OLPC had always been that the entire software stack could be modified, and that users could learn about it and share ideas to make their own platform better.

The Afghan girl risks her life learning how to read and write.

It is the basics of a grade school education which transform and modernize a society.

In these simple things are the roots of independence, power and survival.

That is where your focus needs to be.

The geek builds a machine that reflects his own needs and values and thinks that he has created something universal.

OLPC was interesting, then Microsoft strangled it

[rolfwind]

Just before the netbook explosion too... which started out Linux... until MS squashed that threat... in America at least.

sugar wasn't the problem

[Ephemeriis]

I don't think sugar was the main problem.

Negroponte couldn't seem to make up his mind on the device. First it was supposed to be small, cheap, and completely open-source/user-modifiable. Part of the point was to make the entire platform a learning experience.

Then the hardware specs started changing to make room for Windows... Why exactly? Who knows... Microsoft wanted a piece of the pie, and Negroponte accommodated them.

But then the device wasn't nearly so cheap, and the entire platform wasn't an open learning experience. The cost lost them a few customers... And the lack of openness lost them a few more...

And the marketing? Horrible.

There are plenty of netbooks out there now... Stuff from MSI and Dell and HP... Some ship with Windows, some ship with Linux... They're selling just fine. There's no reason the XO couldn't have been a successful product.

A lot of things combined to kill the XO

[jmorris42]

A lot of things came together to kill XO.

1. Sugar. It still isn't ready for primetime. Building a whole new UI proved a lot harder than designing a laptop for the 3rd world. But worse, Sugar LOOKED like a toy instead of a computer. Basically it was a PR failure even had it been ready to deploy in time.

2. x86. Unless the whole goal was teasing Microsoft into cheap licenses by waving the Penguin flag there was no reason to put an x86 into it. The power problem would have been so much simpler with ARM and the Sugar stack would have ran equally well on it.

3. Failure to understand the customers. The customers were never going to be the children. Neither was it going to be the educators who would have to relearn pretty much everthing to adopt them. The customers were third world kleptocrats.

4. Failure to clearly convey the whole new educational method XO implied and to get any buyin on it. Yes we on /. got it but most MSM reporting on it failed to get it even in the US.

Re:A lot of things combined to kill the XO

[FourthAge]

5. They wanted to create an experimental "mesh" networking system to get all the OLPCs online. This is at least as difficult as making an entirely new UI, perhaps more so.

6. They refused to leverage existing application software, preferring instead to reinvent every wheel for their Sugar system. This was an incredibly bad idea. Compatibility with existing Linux apps should have been a priority to make up for the things that the Sugar apps couldn't initially do.

7. Generally: the project ran on wishful thinking. They set crazily ambitious goals and totally underestimated the time, money and human effort required to achieve them. Intel and Microsoft didn't need to move to kill the project, it would have withered away anyway because of the vast discrepancy between Negroponte's dreams and his abilities. The fact that they tried anyway shows how well the PR machine worked. It was the only aspect of OLPC that was in any way successful. Unless OLPC can claim credit for the netbook revolution, which is doubtful to say the least.

Finally - I expect, if it had been ARM-based, then Microsoft would have been pushing WinCE. However, it is a myth that ARM CPUs are lower power than x86. The instruction set architecture has little or no relevance to power consumption. It's all about the materials used to build the CPU core.

Re:Feature creep killed the XO

[westlake]

If they'd just made the widget, put it into production, and focused on the sales, they would have made a difference.

They built the widget. They put it up for sale. Nothing much happened.

Confirmed sales to date about 1.4 million:

Uruguay 300,000
Peru 290,000
Columbia 165,000
Rwanda 100,000

Sales outside the Americas have been pathetic.

G1G1 167,000 [Distributed in seemingly random 3 to 20K clumps. I can't bring myself to take this part of the program seriously.]

Summary of laptop orders

The demand for Windows came from the third world education minister - the guy who is expected to sign the purchase order for 300,000 units.

Re:Feature creep killed the XO

[QuantumG]

time to market.. heard of it? They hyped up their product then dragged their feet, by the time it was ready, alternatives had been found. This is the reason why the stealth startup was invented.

Re:Feature creep killed the XO

[MacTO]

Feature creep killed the XO, but it had nothing to do with the Microsoft lobby. A few things to consider:

The XO was originally intended to be a replacement for textbooks. It was not intended to serve as a general purpose computer. This feature creep meant that the XO needed additional hardware, higher performance hardware, and experimental hardware. None of this feature creep had anything to do with Microsoft. If you ever used Sugar on the XO, you would be more than aware of this because the performance is substandard. Not only is the system slow due to running interpreted software on a slower CPU, but the software is prone to failure due to there being insufficient memory. (There is no swap space, and two Activities will easily consume all of the RAM.) In other words, and internal desire for more features meant that the hundred dollar laptop would never cost a hundred dollars to build and that the software would never be reliable.

Just to re-iterate, the failure of the XO has positively nothing to do with Microsoft. It did not significantly alter the design of the XO. Nor did it significantly alter the software development process on the XO. About all that it did do was cause some in-fighting between bystanders who would never contribute to the project in the first place.

Re:Feature creep killed the XO

[westlake]

This is the reason why the stealth startup was invented.

The stealth start-up demands stealth funding - usually from someone on the inside.

OLPC needed at minimum a public commitment from an Asian OEM to design and build the display.

No way that was going to happen without hyping sales prospects into the millions and tens of millions of units.

No way that tech wasn't going to surface later in less charitable-minded projects.

Re:Feature creep killed the XO

[QuantumG]

Ya, so: rock OLPC hard place.

Hopefully everyone else learnt the lesson. If you want to compete with Microsoft/Intel/etc you have to strike quickly.

Policy Blunder

[fmorton]

+1 on other remarks that it was not a technical issue

I'd say it was more than marketing, though. It was more along the lines of policy. I think they had an (barely) underlying agenda that this was not about the United States but about trying to bring the world on the same level technically.

I tried to talk with them about buying batches of computers for disadvantaged kids I work with in the US. These kids have no possibility of having their own computer, either. Some of them would also use it as a lighting device since they have no power at times to their house/apartment. But there was no interest on their part.

There are probably 1000s of organizations in the US that would buy 10 or 100 or 500 low-cost computers and would be willing to understand that there isn't much technical support from the seller at those low prices.

They made a really basic miscalculation about the potential audience for the machine.

Yeah, whatever

[coryking]

. Look at it this way - if kids grew up in Africa, South Asia, and South America with XOs, they'd grow up with a clear alternative to buying MS and Intel products when they wanted computing power

OLPC failed, in part, because they went out of the way to please people like you rather then their potential customers. You aren't their customer, you are an arm-chair quarterback. All you have done is added a thick layer of zealotry and politics that have zero place in the business OLPC was in.

I know this isn't going to make me popular in these parts, but at first I was excited when I heard about OLPC until I read their mission statement. The second I read "Free Software" and "GPL", I knew they were horribly unfocused and would eventually fail. The politics of Free Software(tm) have no place in a non-profit that was supposed to put computers in the hands of children. Pushing those goals in parallel with trying to build a computer from scratch, put together an operating system from scratch, putting together a whole new method of education, and *finally* convincing governments to buy said devices from said organization was asking for way to much. Adding "GPL" and all the baggage that goes with it did nothing but bring out the trolls and zealots and stole only shred of focus the company might have had.

Sadly, my prediction was 100% correct. Had they been merely a non-profit trying to put laptops into the hands of kids in developing nations, while they wouldn't have been on slashdot or any other linux rag much, they probably would have had a much better chance to fulfill their mission. A shame, really.

Re:Yeah, whatever

[Improv]

I don't care about "these parts" and I suggest you shouldn't care either.

They wern't trying to just put computers into kids hands - you don't need a nonprofit for that - if you wanted to do that, you could open your checkbook and talk to Dell. The interesting task they were aiming for took most pieces of the puzzle they put together - the degree to which they made their own hardware may have been a mistake, but having a separate branding and having a different software platform for ultra-lowcost computing makes a lot of sense if they're aiming to open the door to community-friendly computing. Windows would not have been a good choice there.

Of course, what they were trying to do was very hard, and I predicted as well that it would not likely make it (particularly as until recently the software stack was rubbish - only the more recent builds have made it nice). The GPL really wasn't a problem, although they did not leverage it well because it was so hard for the OSS community to get the devices to work on them.

It was Negroponte's fault

[mookiemu]

The OLPC is a perfect example of what happens when business challenged intellectuals try to do something. The OLPC was and still is so ahead of it's time and it could have been such a huge success had it's release to the world not been so mismanaged. Every techno-geek I know, including myself, would salivate everytime an article appeared regarding the OLPC. And my mouth still salivates when I read about the OLPC II with dual opposing touchscreens. Had they not made it so difficult to get one, with the limited time offer and the buy-one-get-one-free policy, This thing would have completely taken over the market instead of the eeepc. Then they would have had the funds and the market demand to make them for under $100. And they would have been able to head off the MS threat. So blame Negroponte. His heart was in the right place but he bumbled it with his economic stupidity and as a result lost the window of opportunity to get this remarkable product out into the world. I think there is still a chance with the OLPC II. How many of us here can honestly say we wouldn't pick one up if it were available? I would buy two!

Re:I think what he means

[TheRaven64]

For a project like this, OSS zealotry had no place

I keep seeing this, and you obviously have no clue about the project aims. OSS was the only possible option within the project aims. They wanted countries like India and China, with a large industrial base but relatively poor education, to be able to take the designs, take the software stack, and build their own (improved or direct) copies. This would not have been possible without a software stack that came with redistribution rights. It would not have been possible without completely open designs.