Slashdot Mirror
Negroponte Sees Sugar As OLPC's Biggest Mistake
griffjon writes "In an interview, Nicholas Negroponte claims that the biggest mistake OLPC made was the revolutionary Red Hat-based Sugar desktop environment — instead, he says, they should have built Sugar as an application that ran on a 'vanilla' Linux OS. Some disagree."
Wow, my first thought was he must be on his deathbed and trying to scam his way into heaven. Then, I saw the first name.
I swear to God...I swear to God! That is NOT how you treat your human!
The biggest OLPC mistake was Negroponte.
Things whent downhill the second they started working with MS.
Seriously, they should have stuck to ultra-cheap durable laptops, rather then try to cater to MS's Windows. They lost their focus and thats the end of them.
From the article(writer's words, not Negroponte's):
The "$100 laptop" term was the OLPC marketing failure. If the XO was again called the "Children's Machine", or better yet "the best educational tool for primary school children in the developing world", which isn't as catchy as "$100 laptop" but much more accurate, he would be crowing about multiples of millions of childrens, not just about one.
But calling it something like "Children's Machine" instead of "$100 laptop" might not have given it the chance of catching investor's or public's eye and might have died a death similar to many other types of custom machines. And running it on top of regular distros is not really feasible because the requirements for the OLPC were like 1 GB NAND flash drive and 256 MB. Run Ubuntu on a 8 year old machine now and you will realize that it's exactly very usable, even for web browsing.
This space for rent.
I now have an inexplicable craving for cookies.
Amazing how an interesting and informative comment can be totally ruined by pointless racism...
http://twitter.com/OLDTELEGRAM
Its a shame Sugar didn't turn out to be as popular as it could have been. I know as a kid I used to play with the computer for hours on end, changing settings and playing with QBASIC (gorillas anyone?). By giving children an open source OS to play with (as well as some kind of instruction) they might have really had the opportunity to learn something.
However, in terms of the OLPC goal, they should of gotten on their knees and begged for Windows XP. Giving children all around the world laptops is the more important goal than spreading FOSS, and the lack of a windows environment is what helped its competition grow and crush the project. I remember at least one major sale was blocked because Intel's competing laptop (which was more expensive) had a windows environment. If they would of dual booted sugar, the children would of found it and learned it. If anything else, just to annoy their parents.
shithead who decided microsoft windows simply MUST run on the OLPC in 60 days?? and now his professional evaluation of the open source operating environment he wanted to replace is that it sould have been sitting atop vanilla linux??
whats left? a big bold redmond boilerblate on the case that says "fuck you red-hat" with clippy waving the bird?!
Good people go to bed earlier.
Compatibility is a very, very important feature. And the more complex something is, the more important compatibility becomes!
Every technology has its "API" - the specific interface between it and its environment. And it's very, very, very important to ensure that this "API" is consistent with existing implementations of the technology in order to be successful.
We have many different models of cars, all with their respective features, at price points that range from $2,000 to $200,000 and this is OK because they all have steering wheels, gas/brake pedals for the right foot, and will fit on a standard road.
Take *any* of these basics out of the equation, and you suddenly have a non-starter. The interface between a car and the gas station is but one simple parameter, and yet electric car company startups have come and gone simply because this simple interface breaks.
When looking at an operating system, it's very, very important to maintain compatibility between the operating and applications, sure, but it's also important to maintain compatibility between the operating system and its USERS. It's vexing for users to switch from MacOS to Windows, or from Windows to MacOS, and both have long-established, entrenched interfaces that they are all *very* slow to change. Windows still has it's "X" window kill switch in the top right corner, etc.
With this in mind, it's not a surprise that a whole new graphical interface for a start up caused all kinds of problems. Sure, it's innovative, logical, easily learned, etc. The meta-language Esperanto has all these qualities, yet we all still speak English, with all of its spelling oddities and grammatical exceptions and cruft from its thousand-plus years of history.
I have no problem with your religion until you decide it's reason to deprive others of the truth.
Negroponte sounds like he was the kind of kid that even the geeks stuffed into lockers.
And what a shame, cuz he's always seemed like such a pleasant, down-to-earth, inclusive fellow...
First, advertising it at first as a $100 laptop was a mistake...it's stupid to announce a price before you figure out what your cost structure is going to look like.
Second, and more importantly, the distribution plan was flawed. Their big idea was to sell this thing to the governments of third world countries, despite the fact that most third world countries are led by corrupt governments that have little money, and use what money they do have to grease the palms of the inner circle of the government. Most charitable organizations learned decades ago that trying to get corrupt governments interested in doing something for the interest of their poorest citizens is a recipe for failure.
What they should have done is sold these things to charities that already work in these areas and have knowledge of the difficulties involved and would know where the greatest need is. They should have been dumping these things on charities as fast as they could take them, but instead they were busy trying to get these governments to distribute them, thereby assuring they were only going to be going to countries with relatively stable governments with experience delivering large-scale deployments of things like electrical power to their residents. This means the people that would get them were the ones least likely to benefit.
Third, they didn't do enough to get the American public interested in the project...sure, there were a bunch of stories in tech rags about how cool this was going to be and how no one could get them unless they were a poor person in a third world country, but that was it. This meant the people most likely to have the spare cash to donate to this cause didn't know enough about it, and never had a chance to get their hands on one except through the short-lived "buy one, get one" program.
The Internet is fast becoming what electrical power was 50 years ago: It separates the people who are able to participate in the global economy from those who can't. The so-called "digital divide" has been largely closed in this country, but it remains a huge problem globally. The OLPC program is and has been a great idea with a piss-poor implementation plan.
When your customized system takes 2-3 times as long to boot as Windows on the same hardware, you probably have made a mistake. Maybe next time don't write it all in python.
Klingon programs don't timeshare, they battle for supremacy.
No, the biggest mistake that the OLPC mas that they didn't make cheap, rugged computers. If they had stuck with it, I would imagine they could end up turning a profit by selling cheap computers to the first world. Instead they decided to go with MS and now rather than having a usable, cheap rugged machine their new vision is using (relatively fragile and currently expensive) multi-touch screens for everything.
Taxation is legalized theft, no more, no less.
Sure, but name a desktop environment that would work that well on the OLPC's pathetic specs. Sugar was designed to be primarily single-tasking leaving enough CPU/memory in order to run those tasks well. The temptation to multitask with a traditional DE would end up ruining performance. You have to remember, this thing is more underpowered than the cheap netbooks that will barely surf the web on anything that has Flash.
Taxation is legalized theft, no more, no less.
Develop thicker skin. You don't throw out all the good stuff just because you find one element offensive. If the U.S. government had told Von Braun to fuck off on his rocket ideas just because he used slave labor in his factories and was a SS officer, we wouldn't be celebrating the Apollo landing anniversary today.
SJW: Someone who has run out of real oppression, and has to fake it.
The big mistake OLPC made was Nicholas Negroponte. He's too much into his own self-importance. His thing was dealing at the national leader level and getting himself into the press. What was needed was somebody who knew how to get a low-cost product out the door and sell it in quantity.
The OLPC should have been in a bubble-pack in every Wal-Mart and Walgreens in America, in every souk in the Middle East, and in every market in India, selling at a small profit and dropping in price every three months. But no, Negroponte had to try to make big deals with governments. That might have happened after they actually had the product out there in volume.
The meta-idea of rethinking UI from the ground up, and building something specifically directed at kids, was a wonderful idea. Frankly, the reason I bought a G1G1 machine was that I hoped to experience a fresh and wonderful user interface.
I think Sugar was a sad failure. I don't quite know what to make of the obvious riposte that I don't belong to the target audience. But an awful lot of the official Sugar documentation seemed to me to make too much use of "proof by repeated assertion." A file system organized primarily by recency (the Journal) instead of space (the Apple pre-OS-X Finder) or nested hierarchy (pre-GUI)? Wow, what a strange idea. What a fresh idea. I couldn't imagine how it could work, but all these people said it did, so after giving up on imagining it I paid $400 to experience it. Well, it sure didn't work for me.
And the claim that it works for kids because they "naturally describe what they are doing"--sorry, I just don't believe nine-year-old kids are going to type text tags and descriptions into every Journal entry so that they can find them again. Subject to correction by anyone who's actually watched real nine-year-olds playing with an XO and seen them tag and describe Journal entries, but the last several times I asked this online nobody said they had.
UI design seems to me to have peaked sometime in the early 1980s, when computer companies still needed to seduce laypersons who weren't already trained on computer usage. As "computer literacy" became more and more of a career necessity, computer companies were able to get away with more and more complexity. For me, an important downward turning point occurred when Microsoft violated Apple's UI guidelines, which stated that documents should always re-open with the insertion point positioned where it was when the document was closed--a special instance of the principle that things should stay where you put them. Microsoft couldn't be bothered; with Word, like Sisyphus, you always start with your insertion point once again having rolled down to the bottom of the hill. Other companies, eventually including Apple, followed suit, and this minor but significant point of UI design was lost, along with many others.
A fresh look at UI design is desperately needed. UI design is now in the hands of power-user snobs who revel in their ability to handle complexity. Ordinary people resign themselves to forever feeling that "I'm just a dummy when it comes to computers." The world desperately needs a user interface so simple a child could use it. A pity that Sugar isn't it.
"How to Do Nothing," kids activities, back in print!
``The key to it all is that kids own their machine, so all the admin stuff (networking, power management, etc.) *needs* to work within a consistent, simple GUI.``
That view, and the Sugar UI FWIW, stem from a completely flawed understanding of children. Kids are inherently quick at learning and highly adaptable. Give them a Linux or a Windows UI and they'll thrive, taking that knowledge with them and building on it to adulthood.
What Sugar did was try to lock them in a world of Fisher Price toy simplicity, as if they were intellectually retarded. None of the UI knowledge of Sugar would benefit them later. It thoroughly deserved to fail.
I am in an excellent position to evaluate this issue, having purchased one of the first XO OLPCs through the give-one-get-one (GOGO) program for an 11 year old child, and then obtained an Asus Eee PC netbook running Xandros Linux (a window-ized interface to Debian Linux, along the lines of Ubuntu).
The 11-year old's verdict: thumbs down for Sugar, thumbs-up for Xandros. She gave up with fiddling on the XO after a few weeks, but loves to use the Eee PC. As the network support resource for my household, I can further point out that Sugar shipped with unusable wireless security (WEP only), which some months later was upgraded to WPA, but with this fatal flaw: every time the computer is powered up the user has to reenter the entire passphrase to get wireless access. Since a rather lengthy and obscure passphrase had been previously selected to provide household network security, this was an intolerable nuisance to an 11-year old. And dumbing-down the household security for the convenience of one cheap product is unacceptable to this network support resource.
Perhaps the passphrase remembering problem has since been fixed (since the XO is not used by its target audience any more I am not inclined to upgrade the OS to test it) but it illustrates the fatal problem with the Sugar approach: writing a decent OS is hard work, and taking a quick and dirty stab at it gives a foundation of sand for the whole offering. Absolutely they should have run a solid robust proven OS (Linux) for the system, adding on what ever they felt was needed.,/p>
Starships were meant to fly, Hands up and touch the sky - Nicky Minaj
Before we all jump on the OLPC/Sugar hater bandwagon, let's read the actual quote:
So it sounds like the problem, as Negroponte see it, was the Sugar was written as a pseudo OS/large chunk of user space when it should be acted more like a kiosk application on top of more traditional software to run things like power management.
Sugar it's self, as an interface, still seems like a good idea to me. It's a good idea to give people a simplified user interface for basic tasks. Why should kids in Nigeria or Malaysia have to learn a "traditional" interface? Remember our "traditional" interface exists because of historical storage limits (which for OLPC don't matter, you don't need everything on a 50kb disk, you have hundreds of megabytes free) and user training (they were familiar with CP/M, DOS should do files the same way. They were familiar with DOS, Windows should do files the same way...).
It's very hard for us to switch off of 20-30 years of average people being used to files onto something more like a journal or database, but something like the OLPC was the perfect chance to get computer users who weren't saddled with all that (at least initially).
Sugar: Good idea, implemented in a way which broke the layers of abstractions it should have been built on.
Comment forecast: Bits of genius surrounded by a sea of mediocrity.
I am responsible for a half-dozen different "appliance" server platforms. They all have Linux at the core, but a specialized CLI and GUI (usually web-based) layered on top for administration and maintenance of the box itself, and configuration and monitoring of the application it runs. They are by no stretch of the imagination Unix servers, despite a *nix-like core underneath, where the user can't get at it easily (or at all).
This is standard industry practice in the year 2009, and not a "mistake."
The mistake made wasn't committing to Sugar first, the hardware second - the rush to cram Windows on their boxes was stupid and self-defeating. The OLPC was best categorized as a personal computing appliance rather than a general purpose workstation, and Sugar was and is fantastic for this purpose. By committing to the learning-appliance concept, they could tweak Sugar to run on whatever hardware offered the most bang for its buck. Processors come and go, storage drives obsolete themselves like clockwork... it doesn't matter. The processor isn't the purpose, the RAM is not the point. The point is that the kids have a computer that's easy to learn, rewarding to master, simple to maintain and reliable under all circumstances... and that starts with the interface.
Besides, Apple doesn't have a problem running its interfacer overtop *nix.
Personally, I think their biggest mistake was not selling it to first world consumers. I know a lot of people who would have liked to buy one, but couldn't. This was a fatal mistake since their plan required being able to produce large enough amounts of these to be able to sell them cheaply, and they were turning away the people who were willing and able to buy at the time.
I think this is only a symptom of the biggest mistake, which was a flawed vision of how this project needed to work. Negroponte thought he could swan into the offices of big-time politicians in third-world countries and just talk them into buying these computers en masse, even while he insisted they were not really computers (and could not run Windows) but educational tools. Educational tools? At $100 a head (and climbing)? How many of these countries are investing $100 per student to build schools? What's more, how well has this model ever worked for vaccines, or malaria nets, or cooking stoves? The only way Western countries have managed to bring these things to the poorest people of the world is for independent charities to strap on their boots and go deliver them by hand. Governments are not going to do it for you.
Further, and more to your point, so you put a $199 laptop into the hands of a child of a family that doesn't earn $199 in six months. What then? How much is that kid going to learn about computer programming, open source, and all that other good stuff, when the fields need to be ploughed? One of the main reasons people in third-world countries have lots of kids is that they need them, particularly in areas where people are regularly knocked out of commission by malaria for half the year. So how long is it going to be before that family sells the OLPC?
And then what? Exactly. The OLPC ends up in the hands of ... someone who can afford to buy it. This is Negroponte's real biggest mistake: Denying the basic forces of economics.
If, on the other hand, he had put them into every Wal-Mart -- or screw that, Walgreen's -- and every souk and ever bazaar, in the teeming millions, it might have had a shot. The only way to counteract the economic forces in the poorer regions is for not just the cost, but the value of the device to be low... and the only way to do that is to bump up supply. Keep focused on making the devices virtually ubiquitous, as commonplace as bicycles. In short, the OLPC project needed a lot more people on board and a lot more money backing it. It needed the participation of international charities and it needed to be subsidized by people buying the devices here (at a "novelty" markup, even).
Instead, they went with the "I just need to go shake hands with Nice General Abouda, and he'll help us out" model. Seems like a recipe for failure, to me.
Breakfast served all day!
Missing keys? WTF. People with pocket knives cut them out or something? It's a frigging rubber mat keyboard. How does it loose a key? What kind of trade show was this? A razor blade trade show?
Standard disclaimer, I've used a zillion operating environments, like most /.'ers, with my favorites being VMS and Ubuntu's flavor of GNOME (In MY DAY, we used punch cards, and we had to punch the holes out with a hand-bone of a squirrel we had to catch and skin ourselves. Now get off my lawn!). I also bought a GOGO Sugar machine. My kids hated it, I hated it. The wireless would lose all connectivity after each 24 hours (yes, my DHCP lease was infinite), after taking days of fiddling just to get it to talk to the secure wireless modem. I eventually gave up and got a USB wired ethernet connection. It still would hang randomly, though less often. And the only thing the machine ever did was FireFox. The "documentation" was more missing than real. I chalked all that up to the program being more focused on getting the machines to Peru and Mongolia. But it didn't make the system any less painful to use.
It now sits on the floor in my office, in a dusty heap. If I ever get the motivation I'll put a flavor of real Linux on it, but that's pretty unlikely.
If it had come out with a semi-standard Linux interface I think it would have stood a better chance of success, if only to me. How long did it take to develop the environment and apps when the resources could just as well have been devoted to making a lean and mean distro to fit the hardware? Poor resource planning, it seems.
In the immortal words of Tom Lehrer:
Gather round while I sing you of Wernher von Braun,
A man whose allegiance
Is ruled by expedience.
Call him a Nazi, he won't even frown.
"Ha, Nazi Schmazi," says Wernher von Braun.
Don't say that he's hypocritical,
Say rather that he's apolitical.
"Once the rockets are up, who cares where they come down?
That's not my department," says Wernher von Braun.
You are not a brain: http://books.google.com/books?id=2oV61CeDx-YC
They didn't just lose focus, they lost a lot of goodwill by working with MS.
The push for XP came from the education minister - the guy who is expected to sign a purchase order for 100,000 units.
This was a fatal mistake since their plan required being able to produce large enough amounts of these to be able to sell them cheaply, and they were turning away the people who were willing and able to buy at the time.
The XO-1 was something of a cross between an e-book reader and a netbook - when neither product was clearly defined or particularly economical to produce.
The first to dive off the pier-
usually misses the deeps, hits his head on a rock and drowns.
The netbook may still lack a clearly defined market: Many netbook buyers aren't happy [June 21]
It wouldn't be entirely unfair to describe sales of the Linux netbook as "a flash in the pan."
OLPC needed to sell millions of units each year to avoid being lapped by its commercial competitors. I don't think that was ever going to happen.
That's not a good comparison to the post that sparked his comment.
Von Braun offered more than comments. When a person's entire topical output is a written comment on a message board, then overt racism in the comment is cause to ignore the whole thing.
Simply put, a comment on slashdot is not *important* enough to override the derision such outright racism deserves.
That's my opinion, yours might differ -- but I think that overlooking such blatant racism is tantamount to approving of the racism.
"Trolls they were, but filled with the evil will of their master: a fell race..." -- J.R.R. Tolkien on Olog-hai
Actually, no. Xerox PARC established what a computer interface should look like. Apple took from Xerox & improved it. Microsoft took from Apple &... well... became more commercially successful for reasons mostly of a preexisting monopoly, namely MS-DOS.
To make a really bad car analogy, bad marketing didn't kill GM, in fact marketing kept GM on life support for decades *in spite of* their crappy cars.
The XO as delivered during the first G1G1 (and to country projects during that period) was nearly unusable. Hardware functionality was great, but the software didn't measure up at all, in several respects. Software was still experimenting in blue sky when they needed to be delivering on goals.
First, it didn't perform acceptably within the severely limited RAM (severe for non-memory-footprint-optimized software). Most likely running the same software on the XO gen-1.5 refresh will fix that level of performance issues.
Second, the software was not complete, in that promised features were not yet implemented. Some were important in the real world, like the advanced power saving. Things crashed a lot.
Third, the software implemented experimental ideas like the Journal, which were under revision without taking end-user feedback into serious consideration.
[And Fourth, it was damned hard for willing FOSS volunteers to contribute meaningfully. The build you could easily download and run and report bugs on was far obsolete from what the developers were running, and getting in sync with them was tough.]
From what I could see on the developer mailing lists, they had a severe cat-herding problem. Too many smart people going in whatever direction their creativity and youthful self-confidence took them, rather than getting all hands on deck to recognize the shortcomings and get them fixed *now*. I think they undertook this project in the spirit of constructivism, when they needed to switch gears and deliver a product. Too much green hat, not enough blue hat.
Those are my opinions as a fan of the project.
It's a shame for everyone who read deeper into this, because the whole pretext of this story is based on an over approximation.
Negroponte said this:
[T]he biggest mistake was not having Sugar run as an application
He in no shape or form said Sugar was a mistake! He is talking about the implementation, you fools! It is like saying the automobile was a mistake, when the inventor just said he should have used a cleaner engine.
The person disagreeing with the words they put in Negroponte's mouth says:
Sugar was not a mistake, it is one of the defining aspects of the XO laptop, and saved it from even more unfavorable comparisons to traditional laptops and accusations of being underpowered.
Right. So in what way are you disagreeing with the claim that Sugar should have been more modular, the system architecture should have been simpler, and that Sugar could have been more interpolatable with other systems?