A View From Inside the OLPC Project

icknay writes "Here's an interesting rant on the OLPC from someone who worked there, including: 'The core mistake of the present Sugar approach is that it couples phenomenally powerful ideas about learning — that it should be shared, collaborative, peer to peer, and open — with the notion that these ideas must come presented in an entirely new graphical paradigm. We reject this coupling as untenable. Choosing to reinvent the desktop UI paradigm means we are spending our extremely over-constrained resources fighting graphical interfaces, not developing better tools for learning.' I have an OLPC, and the OS itself seems quite unfinished. I buy the argument that it would be better to focus on Sugar as educational software, and let it run on Linux, Windows, whatever."

Comments

We are not in the dark.

[Odder]

There's a lot of spin and intentional ignorance here and it spills out best when he says this:

we don't know that laptop recipients will benefit from fixing software on their laptops. Indeed, I bet they'd largely prefer the damn software works and doesn't need fixing. While we think and even hope that constructionist principles, as embodied in the free software culture, are helpful to education, presenting the hopes as rooted in fact is simply deceitful.

The project in Sengal was not the only place non free software has bombed in education - it's bombed everywhere, not due to "intense competition" but to greed and planned obsolescence. Non free software is mostly designed for business, not education. What little non free software there is is quickly obsoleted by the upgrade treadmill and must be replaced at great expense. The dominant OS has been even worse from the very beginning with poor security - from macro viruses on floppies to today's modern botnets. The net result is that only the richest of schools has been able to afford a good ratio of computers to children and they do little more with them than write papers. They lacked free libraries, text books and other useful references until these things showed up on line. Schools like MIT did better because they helped themselves, in part with free software. The non free way has been an unmitigated dissaster and should not be pushed onto anyone else.

Re:We are not in the dark.

[tsm_sf]

Why do so many people have such a hard-on for this guy? He has several previously identified accounts, so just tag them as 'foe' and be done with it.

((btw, if you feel that you are "educating the masses" in regards to general douchebaggery, please be advised that we've seen it all))

Re:We are not in the dark.

[Lars512]

Presumably you're unable to deploy 40,000 laptops in Peru without backing or support from either other NGOs or the Peruvian government. He simply said that there was only one staff member dedicated to deployment. Perhaps their job was to coordinate a much larger pool of non-OLPC staff and volunteers. Without knowing the scale of OLPC staffing or the support from Peru they got, it's difficult to determine if this situation was really so poorly managed.

Re:We are not in the dark.

[catbertscousin]

What the hell is "intentional ignorance"? Choosing not to RTFA.

Re:We are not in the dark.

[hey!]

But the Rudd government is still young, and we are yet to discover the riddles and traps of their policies.

I'm all for cynicism, but if you go too far you end up going full circle, approaching naive idealism from, as it were, the other side. It appears that you think that governments should operate more effectively other human institutions.

People are fallible, and inconsistent. In fact, given that people are fallible, it's a good thing they're inconsistent. But things are never quite what you expected them to be; over time you are forced to compromise so many times that the initial coherence of your plans starts to fall apart.

That's why the best way of managing human enterprises is to make them mortal. The Grim Reaper removes the deadwood and lets the light in for new growth. This notion is built into capitalism. Businesses fail all the time. Business that are "too big to fail" incorporate extinction into themselves, changing leadership, spinning off divisions, dropping products and initiatives all the time.

That's why crony capitalism stinks. It takes extinction off the table. Businesses, like large defense contractors, that are treated as national assets are a mix of the worst elements of government and private enterprise.

The real value of democracy is not to ensuring good government. That's humanly impossible. The real value of democracy is removing bad government. The only advantage of democracy over any other philosophy of government is that it makes extinguishing a bad regime routine and in social terms relatively cheap. People living under other kinds of systems can change regimes, but at the risk of being killed, imprisoned or exiled.

What we are seeing in the OLPC program is a questioning of some of its initial assumptions. That's inevitable. It's also unfortunate that MS used it's marketing pull to force them to confront the Windows issue,because given the mission of the organization, that's a side show. And, as confronting issues in these cases always does, it's undermined other assumptions of the project's initial vision, like creating a new desktop paradigm. It would have been better for the project to have had more of a chance to make its case before it faced political pressures.

In defense of the Sugar concept, it's sometimes easier to do more than to do less. It's easier to keep the creative juices flowing, to attract interest and support, when you are doing something revolutionary than when you are doing something incremental, even if the incremental changes have revolutionary effects.

But that's only when things are going well. When the project's assumptions are undermined, then bold design philosophies become a liability, sapping support and energy by being one more thing that people argue over.

The problem with OLPC and Windows

[mlwmohawk]

Forget Sugar, yea its great and all, but the point of the OLPC is learning. Learning requires freedom.

Windows is not "free," and I don't mean price, and I mean freedom. Putting Windows on OLPC is nothing more than a marketing move by Microsoft. Not to help kids, but to ensure they become customers. Not giving them books, selling the subscriptions. Not teaching them to farm, but making them sharecroppers.

Re:The problem with OLPC and Windows

[initialE]

Learning requires freedom. At the risk of being flamebait, exactly how does learning require freedom? Children learn from their parents - the most autocratic system in the world is the family structure, especially in the formative phases. Yes, freedom is a good thing to have, but it's not going to benefit people if all they learn to do is use an obscure system that doesn't do anything the way they do it out in the business world.

Re:The problem with OLPC and Windows

[yomegaman]

When I grew up we had no idea what free software was, all we had were our Apple II's, C64's, etc, that were pretty much 100% proprietary. Yet, we somehow learned about computers by reading books and writing our own programs in the cruddy BASIC interpreters they came with. A kid with XP, Java/Python/what-have-you, and the Internet is a million times better off than we were. I swear, some of you people act like it's a tragedy if someone grows up not knowing Bourne shell scripting. The platform you learn on isn't that important, as long as you are learning the concepts.

PS: Besides, you can use a computer to learn about things other than the computer itself, right?

Re:The problem with OLPC and Windows

[pembo13]

platform you learn on isn't that important Microsoft certainly doesn't seem to share that sentiment

Re:The problem with OLPC and Windows

[Karma Sink]

But what help do you get out of it just because it's OSS? I don't see any solid argument there at all. Simply because your software is open source, you are learning more?

It's the 'distro' that matters here. It's how the software is aimed at them and how they help the users to learn to use that software. I don't see how that is easier or harder on OSS or proprietary software.

Re:The problem with OLPC and Windows

[DerekLyons]

You can't do that with any modern computer. You can't learn from watching a multi-layer motherboard where you can't find out what connects to what in what fashion without a multi-thousand-dollar lab and a high-res X-ray machine. You can't just look up what a modern thousand-leg GPU does the way you could with a 74LS74. There are no books on that. You can't cut a trace and rewire something, not anymore.

Nor do you need to do all those fancy things. The vast majority of modern programmers got their start programming, not hacking hardware I wager.
 
 

Different times require different tools. Open source is probably the only way to see what happens in a computer these days.

Sorry, but you are comparing apples to oranges - because you start with hardware and then swap to software. Along the way, you somehow missed that a whole generation of programmers got their start on proprietary computers (Apple, C64, IBM/clones) with a crappy language and tools (BASIC) and a handful of books and magazine (no interwebs). Today, one can start with Python/Java/Ruby/whatever and the internet and be light years ahead of where we all started out.

middle ground

[nguy]

I think plopping a full-blown Gnome or KDE desktop on the OLPC would be a mistake: those desktops work poorly on small screens, and they are incredibly obscure for new users (although no more obscure than Windows and Macintosh).

I think there's a middle ground, though: reuse the Gnome desktop infrastructure but replace the window manager with something simpler that prevents the usual beginner mistakes (losing windows behind each other, moving windows off-screen, etc.).

As for Windows on OLPC, I don't get it. Even if you run Windows+Sugar on the OLPC, you won't be able to install commercial software or commercial drivers with it, Windows books won't apply, and realistically you won't be able to run Microsoft's development tools on the OLPC either. But you will alienate lots of OLPC contributors, and you'll saddle yourself with an OS over which OLPC has no control, and Microsoft secretly probably just wants to kill the whole project anyway.

Re:middle ground

[Alex Belits]

think plopping a full-blown Gnome or KDE desktop on the OLPC would be a mistake: those desktops work poorly on small screens, and they are incredibly obscure for new users (although no more obscure than Windows and Macintosh). I already have a version of Ubuntu with Xfce that has default configuration designed to be usable on those laptops -- it's my development/mobile-device configuration. I even went as far as re-painting icons from Human theme green, so they don't clash with colors usable on a white-and-green laptop. The goal was to:

1. Port a Debian-based distribution with good hardware support, development and "mainstream" connectivity tools.
2. Make configuration suitable for a person who is accustomed to "traditional" windowing systems.
3. Demonstrate that if Windows on OLPC laptops is addressing a problem, that problem is already solved better by using existing free software.

So far I find that laptop perfectly usable -- in fact, for some things it ended up being better because slow Flash annoyed me enough to add a script, mplayer configuration and rebuilt clive package, so Youtube works in fullscreen without glitches. On my regular laptop I did not bother, and just accepted that I have to use Flash plugin with it craptastic performance on videos.

Re:middle ground

[nguy]

I already have a version of Ubuntu with Xfce that has default configuration designed to be usable on those laptops

I don't think XFCE window management is any better for these kinds of screens than Gnome, and I think the XFCE dock and toolbars are considerably worse.

2. Make configuration suitable for a person who is accustomed to "traditional" windowing systems.

But that's not the goal of OLPC. The goal of OLPC is to make a system usable for kids and non-experts. Kids and non-experts have real trouble with Windows/Mac/XFCE/Gnome-style window management. Sugar may not be the right answer to this problem, but none of the traditional desktop environments are suitable.

Amateurs talk strategy...

[edremy]

Professionals talk logistics.

It's an old military saying, and it's right. By far the most damning bits in his article don't deal with Sugar, Windows or anything else- they deal with the utter and total lack of planning on the part of the deployment folks. (Err, folk) The fact that they had virtually no plan, no infrastructure and no supply chain management indicates to me that they were simply not living in the real world- any Army 2LT could have sat down with them and explained how they were about to fail. How you get to a point where you have a quarter of a million pieces of hardware sitting around with no coherent way to get them to the people who actually need them is beyond me. Why didn't they hire a pile of old brigade S4s? You know, folks who actually have experience getting stuff to people out in the middle of nowhere?

I've been tremendously disappointed by the entire project- the goals were wonderful, the hardware ended up pretty nice, the software has ended up pretty meh, but the overall project seems to be run by pie-in-the-sky idealists, Open Source fanatics and others for whom the real world is a place they only visit from time to time.

Learning inhibitors

[mudshark]

This guy is a bit unhinged and it harms his case. He really goes nonlinear about 12 paragraphs down when after tries to rip RMS a new one and says

If proprietary software is half as good as free software at aiding children's learning, you're damn right it makes the world a better place to get the software out to children. Hell, if it doesn't actively inhibit learning, it makes the world a better place.

Well, I respectfully submit that the worldview favored by Microsoft actively inhibits learning. As a blindingly mundane example: Make an OS (Windows) which uses filename extensions to divine metadata about certain files (bad, but we'll let that slide for the moment). Next, release a version of said OS which has a default UI setting to hide these filename extensions from the user. This very demonstrably inhibits learning -- even the casual user picks up fairly quickly on things like ".txt" and ".exe" -- and gives people a distorted picture due to the missing information. That, in turn, increases confusion (why are there 4 things called "Setup" in this folder, why do they have different icons and which one do I click?) and paves the way for some of the the crudest exploits (somebadvirus.doc.exe) simply by dumbing down the user. Not only has the prevailing approach by the monopoly software vendor actively inhibited learning, but the net result of that has been several iterations of malware which Just Didn't Need To Happen.

How can you develop a culture of innovation when you promote a mindset which discourages tinkering? Sorry, but in this case half a loaf is worse than no loaf at all. People like Krsti should at least be able to notice this bias in proprietary operating systems and applications. He makes enough reasonable points that it's even more important not to let him off the hook for something like this.

Re:Learning inhibitors

[Anonymous Coward]

I'll agree that hiding extensions is pretty stupid and confusing, but I can't go along with the theory that it somehow causes brain damage to users, lowers their IQ, and gives them a learning disability. It's not really THAT bad.

Re:Learning inhibitors

[urcreepyneighbor]

How can you develop a culture of innovation when you promote a mindset which discourages tinkering? Look, we're talking about developing a cheap, little, rugged laptop so little M'beka can learn fractions, read Wikipedia articles, and IM his friends about tonight's home.

This "culture of innovation" bullshit - and the rest of the FOSS ideology - is the anchor around the OLPC's neck.

Let M'beka learn.

Enlighten me

[Jooly Rodney]

The cost of developing it aside, what is the problem with having the ideas "presented in an entirely new graphical paradigm," when you're giving the machines to communities in which the per capita rate of computer ownership is practically nil?

Re:Pretty much.

[nuzak]

Did you even bother to read the rest of the article? He doesn't even want XP on the OLPC. What he wants is some focus on the application usability in order to further constructivist learning, regardless of the operating system underneath. The damn thing ships with Squeak, the apps are written in python, and they SHOULD manage to run on any platform.

I think most people read about a page in, then rushed back to slashdot to muster their defense of Free Software and Fight The Good Fight, and well, pretty much proved his point: OLPC's mission is being lost by people who care more about meta-issues than either the learning mission (enabled by the software, not really the kernel) or the ongoing viability of the project itself (deployments need support!)

Peru may soon be stuck with 40,000 doorstops. Maybe I'll go take a look at Sugar and see if any of the ideas are worth lifting for a groove-like P2P network.

Re:The jury is still out

[1u3hr]

You could also argue for just using the cheapest Windows or Ubuntu notebooks

I'd argue with that. Half of them would be dead in a month. 90% by the end of a year, in a village environment.

The OLPCs are rugged, pretty waterproof, batteries (a VITAL point) cheap and much longer life than standard laptops.

Comment removed

[account_deleted]

Comment removed based on user account deletion

Re:I'm not buying it...

[Lonewolf666]

This is also why to the chagrin of many geeks, the desktop metaphor just won't go away. It works, and it works well.

True.
I'd like to add that the occasional annoyance is usually due to poor implementation. Examples include
-inconsistent GUI design in general, where you learn some principle just to find out it is NOT applicable to all of the application
-Windows Explorer failing to update itself after you moved around some files
-Focus stealing by popups (Windows again)

All of these are not a fundamental failure of the desktop metaphor, but individual failures to apply it well.

Re:Uh, isn't that the whole point?

[chris_sawtell]

The whole point of XP on XO is that Microsoft cannot stand up commercially if it ever becomes generally accepted knowledge that there are other O/Ss for small computers which work just as well, if not better than Microsoft products. This is what really gives Bill Gates and Steve Balmer serious laundry problems during the day and horrendous dreams at night. They just cannot allow that to happen.

Where the OLPC people are really in la-la land is thinking that the pupils and their teachers are going to be able to produce the course/learning software modules for themselves. The first world has failed spectacularly in that department, I'd really love to think the third world is going to be able to show up the first world as a bunch of ninnies in this regard, but I fear not.

After watching my son's schools futzing around with both desktop and laptop machines, in my not so humble opinion, laptops in primary schools are a complete waste of time, money and effort, and of very questionable value in secondary ones. Useful for teachers to keep records and to produce teaching materials, but for the pupil's use, no.

It matters not one jot who wrote either the GUI or the underlying O/S, because that's al hidden under the course-ware, which is what counts.

Re:I stopped reading after this

[Karma Sink]

And if you can't see the difference between frustratingly banging your head against a keyboard to get Linux to run properly and learning something new every day, then you're not really seeing with very clear eyes, I don't think.

Frustrating the end user so that they think that the system is nothing but a frustrating, annoying piece of crap is not a very good way to get people to work with it. Especially not the people who give up quickly because they would rather be trying to figure out how to get potable water instead of how to compile a web browser. :(

Re:The jury is still out

[anothy]

amen. it's amazing to me that so many people here on slashdot have such a poor grasp of fundamental engineering principles (it probably shouldn't, but it does). the XO is explicitly designed for kids in developing countries. its closest "competitors" are simply not - they're shrunken laptops. it's not a question of better or worse in some abstract absolutist sense, it's a question of engineering your solution to fit your problem.