Negroponte on OLPC's commitment to Sugar

[Posted April 23, 2008 by corbet]

From:		Nicholas Negroponte <nn-AT-media.mit.edu>
To:		devel-AT-laptop.org, sugar-AT-laptop.org, community-news-AT-laptop.org
Subject:		on Sugar
Date:		Wed, 23 Apr 2008 12:06:32 -0400
Message-ID:		<6.2.1.2.2.20080423120536.0618d150@hq.media.mit.edu>


People keep asking me:

Yes, OLPC s commitment to Sugar has changed. It is now larger, not smaller.
Contrary to inferences drawn by Walter's departure, the press and venerable
sources such as OLPC News, we are scaling Sugar up, not down. Let me explain.
 
Sugar is a very good idea, less than perfectly executed. I attribute our
weakness to unrealistic development goals and practices. Our mission has never
changed. It has been to bring connected laptops for learning to children in the
poorest and most remote locations of the world. Our mission has never been to
advocate the perfect learning model or pure Open Source. I believe the best
educational tool is constructionism and the best software development method is
Open Source. In some cases those are best achieved like the Trojan Horse,
versus direct confrontation or isolating ourselves with perfection. Remember
the expression: perfection is the enemy of good. We need to reach the most
children possible and leverage them as the agents of change. It makes no sense
for us to search for the perfect learning model.
 
For this reason, Sugar needs a wider basis, to run on more Linux platforms and
to run under Windows. We have been engaged in discussions with Microsoft for
several months, to explore a dual boot version of the XO. Some of you have seen
what Microsoft developed on their own for the XO. It works well and now needs
Sugar on top of it (so to speak).
 
As a non-profit, humanitarian organization, OLPC has a unique position, from
which it can change the world for children and learning. Laptop makers rushing
into the low-end marketplace is a perfect example of success of one kind.
Another will be what kids do outside school and with other kids around the
world. A third is what we do.
 
We are not a business, but need to be more business-like: meet schedules,
manage expectations and fulfill promises. To do that, we need to hire more
developers, work more together and spend less time arguing. Because of public
attention, anything we say will be quoted out of context. We can only speak
with our actions and those are only one: a reliable and ubiquitous Sugar. That
includes being more collaborative engineers ourselves and engaging the
community better. Our limitations are not financial, but identifying the
required human resources and resolve to do so.
 
What is in front of us is an opportunity for big change. Sugar is at the core
of it. To pretend otherwise would be a joke. That said, Sugar needs to be
disentangled. I keep using the omelet analogy, claiming it needs to be a fried
egg, with distinct yoke and white, rather than having the UI, collaborative
tools, power management and radios merge into one amorphous blob. Otherwise, it
is impossible to debug and will be limited to the small, albeit growing, world
of the XO hardware platform.
 
As we reach out to engage a wider community, some purism has to morph into
pragmatism. To suggest that this forsakes Open Source or redirects our mission
is absurd. Kids will be the agents of change and our job is to reach the most
of them. That is not just selling laptops, but making Sugar as robust and
widely available as possible.

Nicholas

_______________________________________________
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel

(Log in to post comments)

Posted Apr 23, 2008 18:40 UTC (Wed) by gvy (guest, #11981) [Link]

> some purism has to morph into pragmatism
But not some pragmatism into some blindsightedness.

Wonder if he would take money from his political opponents while in dire state (if offered).

Negroponte on OLPC's commitment to Sugar

Posted Apr 23, 2008 18:54 UTC (Wed) by JoeBuck (subscriber, #2330) [Link]

I'd be happy to see Sugar running on top of more platforms. I got an XO for my daughter, but many of the coolest apps require another machine running Sugar; it would be cool to be able to run a second instance from an ordinary Linux laptop.

And I do need to confess our impurity: my 10-year-old daughter happily runs Linux, but she would be much less happy without the proprietary Flash plugin that lets her play large numbers of games, or the non-free codecs that allow videos to play. She plays with the XO some, but Flash has been an issue for her.

swfdec is improving, but chasing a moving target is going to be tough.

Negroponte on OLPC's commitment to Sugar

Posted Apr 23, 2008 21:40 UTC (Wed) by DonDiego (guest, #24141) [Link]

What non-free codecs does your daughter require to play videos?

Negroponte on OLPC's commitment to Sugar

Posted Apr 24, 2008 12:30 UTC (Thu) by zooko (guest, #2589) [Link]

My two sons were given OLPCs, but they don't use them for much, yet.  They request to use my
Linux workstation (for playing and editing Dungeon Crawl) and my Windows laptop (for
Headsprout, which has been gloriously successful at helping both boys learn to read -- our
three year old is reading pretty well now), and my seven year old requests that I fix the
hardware problem in his Linux workstation so that he can play Dominions 3.

So I infer that I am going to have to install Flash player on the little one's OLPC so that he
will use that to play Headsprout instead of my Windows laptop so that I can use the latter for
work.

Negroponte on OLPC's commitment to Sugar

Posted Apr 24, 2008 23:01 UTC (Thu) by cpeterso (guest, #305) [Link]

Installing the Flash Player on the OLPC is very easy:

su -l
wget flash90124.notlong.com
rpm -i flash-plugin-9.0.124.0-release.i386.rpm
exit

as per http://wiki.laptop.org/go/Adobe_Flash

OLPC Jumps Shark

Posted Apr 23, 2008 19:31 UTC (Wed) by jhoger (guest, #33302) [Link]

My summary of this mess: NN, wants OLPC to be a software platform and user interface, not a
laptop. He's decoupling the software from the hardware.

I have no idea what sort of muddled thinking got him to Micrsoft as the right partner for
that. Microsoft has ZERO interest in software platforms that conceal the OS. Hmm... Netscape
was going to be the new software platform. What happened to them? Java was to be the
run-anywhere platform. Microsoft tried to squash that too.

OLPC leadership just jumped the shark.

The best that can be had at this point, short of NN coming to Jesus, is fork sugar as an
actual, non-captive FOSS development project. Hope and pray that OLPC XO-1 was enough to
encourage the marketplace to continue creating laptops in the vein of eeePC, Everex Cloudbook,
etc. but with high enough battery life, mesh networking, and screen brightness to meet the
requirements.

OLPC could be solely concerned then with sales and marketing of OEM laptops integrated with
Sugar as the UI. If they are smart they will us Linux as the substrate. Windows? That's plain
insanity unless Microsoft makes the cost for the OS component disappear.

-- John.

Negroponte on OLPC's commitment to Sugar

Posted Apr 23, 2008 20:34 UTC (Wed) by schwaang (guest, #19827) [Link]

I haven't yet seen a thoughtful analysis of what is "wrong" with the current XO+software.  It
seems clear that things aren't going as well as had been hoped, but is WindowsXP-Lite the
answer to any of the problems?

My own experience with the XO as a G1G1 donor (and prospective FOSS developer) is that Sugar
needs a lot of polish under the direction of a benevolent dictator who can herd the cats in
the right direction, with lots of input from actual educators and educatees.  Windows makes no
impact there.

Maybe in some countries sales of the XO would increase if Windows were available.  But for
OLPC to stay true to its mission, it *must* support sugar.

Is sales competition the issue, or just what issues are being addressed by this move at OLPC,
in Negroponte's own opinion I wonder?

Negroponte on OLPC's commitment to Sugar

Posted Apr 24, 2008 3:52 UTC (Thu) by bronson (guest, #4806) [Link]

> it *must* support sugar.

Why?  I own a G1G1 too.  I love the hardware but I find Sugar to be fairly unusable.  It's
outrageously slow, funky, inconsistent, limited, and ugly.  It makes for an OK demo but, if
the six months are any indication, it's going to take years for it to become a good, solid,
useful front-end.

Why didn't the OLPC guys adapt more existing software rather than writing so much from
scratch?  And why are they so Python-centric?  A Telepathy+XFCE+Webkit front-end would
probably be fairly doable.  The mobile phone guys have been doing cool things; follow their
leads.

Given the huge delays and limited functionality, it seems pretty clear that they bit off a lot
more than they can chew.  And unless 20 programmers suddenly materialize to work full-time on
it, I don't see things getting better any time soon.  It breaks the heart.

Said from the snidelines, of course.  I'm just an interested observer.

Negroponte on OLPC's commitment to Sugar

Posted Apr 24, 2008 19:18 UTC (Thu) by schwaang (guest, #19827) [Link]

I don't disagree that Sugar is limited at this point, and there is a lot of unrealized
potential. Its lack of polish and clunkiness was my biggest disappointment with the XO when I
got my G1G1. I'm starting to appreciate its possibilities more and more.

The reason they must support Sugar, IMHO, is that education is the whole point of this
project. Anyone *could* develop activities for this system more easily than for Windows
(without Sugar), and a kid really can program in Python. The whole system is laid bare.

In any case, the choice of software stacks isn't the fundamental problem, I think the effort
must have been under-funded and, as I said, maybe needed more focus to its effort to get the
thing polished for the needs of its end-users. (I.e. "Benevolent dictator with usability
input for educators and educatees").

>And unless 20 programmers suddenly materialize to work full-time on
it, I don't see things getting better any time soon.

Lots of programmers (like myself) only just got their XOs in the last few weeks via G1G1. If
I can find a way to contribute I'll gladly do it. But I'm a little disconcerted to see the
Sugar folks admirably charging forward without the kind of managerial direction and user
feedback that success demands. With a Steve Jobs laser-like focus yoking effort to vision,
Sugar could rock.

But then, I'm also just yapping from the "sidelines" at this point ;)

There is one other thing I've come to realize about the XO. In the absence of an XO-centered
curriculum, there's a limit to what any child is likely to accomplish with one out-of-the-box.
But in the presence of such a curriculum (given the needed activities), today's Sugar isn't
going to be an impediment to the learning mission.

So it comes down to how they get deployed, and whether the adopting country organizations are
able to stuff some locally useful activities on them that allow teachers to teach and children
to learn.

Negroponte on OLPC's commitment to Sugar

Posted Apr 25, 2008 18:31 UTC (Fri) by amk (subscriber, #19) [Link]

Last month I went to an OLPC meetup and helped someone who wanted to modify the Record
activity on their OLPC.  Using Python for Sugar activities doesn't mean students will be
editing the software on their machines, because the code is still complicated.  

Activities are GTk+ applications, so you need to know about the GTk+ signals-and-slots
mechanism, what widgets are available and their properties, etc. plus a collection of
Sugar-specific modules.  Python is better than C for this, but modifying activities will still
be pretty hard, out of reach for all but the most determined students (and many adults!).

I'd say they got exactly what they wanted from us

Posted Apr 23, 2008 20:53 UTC (Wed) by jmorris42 (guest, #2203) [Link]

My guess OLPC got from us (us being the OS/FS crowd) exactly what they wanted. Which was
exactly what Asus got. Microsoft's attention.

Both wanted XP really really cheap. Both knew that the most reliable way to get it was to
wave the Linux flag and prove viability. As long as OLPC looked like vaporware Microsoft was
perfectly content to allow RedHat and a bunch of idealistic volunteers to waste their time
developing software to run on it. Once they shipped working hardware and showed every sign of
shipping a lot of units Microsoft had no choice but to offer up XP to keep their monopoly
position. OLPC knew this would happen and almost certainly planned on this outcome from day
one. Had they really planned on staying with the Penguin they would have used an ARM based
one chip solution and saved a lot on the 'ol power budget. The ONLY[1] reason to insist on
x86 compatibility is keeping the door open to Windows.

Note that most of the same applies to Asus except they were producing in partnership with
Intel as a flagship for their new low power chipsets so using an ARM wasn't an option. From
day one they were including all of the drivers for XP with each unit with the expectation
many/most would be reloaded after purchase. And note that just as soon as they demonstrated
volume sales[2] they used that to negotiate a really sweet deal for XP. I kinda doubt even
Dell got prices on XP so low they could sell Windows and Linux for the exact same price except
they toss in 8GB of flash as a bone to the poor saps who buy soon to be abandoned Linux
version.

[1] Remember that OLPC lacks the excuse of needing the x86 only Flash plugin since they don't
ship it.

[2] To be fair, the original plan was to retail for $199. When that didn't work it probably
made business sense to rethink the Linux decision since $500 machines do have the margin to
cover a Windows license.

I'd say they got exactly what they wanted from us

Posted Apr 23, 2008 23:56 UTC (Wed) by accensi (guest, #11754) [Link]

Microsoft only wants to play on mature markets. That is the main point. Asus need to open the
market, used Linux and now made a deal with Microsoft (or was forced), where price should be
the same for both versions, even if necessary to put a little more hardware in the Linux
version. Almost fair, probably to avoid regulation issues.

This month in the Brazilian edition (#41 - April '08) of Linux Magazine (the German branch,
not the American one), with an interview with the Strategies Manager of MS Brazil, where he
presents this "strategy" in full terms. The article is only available online for subscribers
of LM.

I'd say they got exactly what they wanted from us

Posted Apr 24, 2008 1:33 UTC (Thu) by midg3t (guest, #30998) [Link]

Could you mention or quote any of the key points of the article?

I'd say they got exactly what they wanted from us

Posted Apr 24, 2008 14:41 UTC (Thu) by accensi (guest, #11754) [Link]

Links is http://linuxmagazine.com.br/article/para_facilitar_a_inte...

As I said, it is restricted to subscribers. I will try later night to do some summary from the
copy at home.

I'd say they got exactly what they wanted from us

Posted Apr 25, 2008 11:30 UTC (Fri) by accensi (guest, #11754) [Link]

Found an online link:
http://linuxmagazine.com.br/noticia/entrevista_ms_brasil
You can use Google Translate for get an English version

Go to the answer of the 3rd question.

I'd say they got exactly what they wanted from us

Posted Apr 24, 2008 0:00 UTC (Thu) by salimma (subscriber, #34460) [Link]

[1] Remember that OLPC lacks the excuse of needing the x86 only Flash plugin since they don't
ship it.

Indeed. It's not that Adobe is not amenable to porting the Flash plugin to non-x86
architectures; the Nokia N8x0 tablets are Linux/ARM-based and sports a proprietary Flash
plugin. Not the fastest thing on earth, but considering how buggy the Linux/x86 version is,
not bad.

I'd say they got exactly what they wanted from us

Posted Apr 24, 2008 1:21 UTC (Thu) by gdt (subscriber, #6284) [Link]

To be fair, the original plan was to retail for $199. When that didn't work it probably made business sense to rethink the Linux decision since $500 machines do have the margin to cover a Windows license.

Confuses costs and prices. Asus have the hot machine of the year. They're obviously going to sell it for as high a price as the market will bear.

As for hiding the Windows costs in the margin, they've gone the other way and reduced the hardware costs of the Eee 900 running Windows. They're retaining that fat margin for as long as they can (ie, until MSI and HP actually ship a competitor rather than talk about shipping a competitor).

That says something very interesting about the low cost of their Linux, even with its heavy modifications.

And note that just as soon as they demonstrated volume sales[2] they used that to negotiate a really sweet deal for XP.

Based on the differences in hardware, the Microsoft license cost looks to be around US$40. That isn't a sweet deal, in fact it's so high that it looks like Microsoft are penalising Asus. I wonder if Asus rejected a sweet deal from Microsoft which proposed dropping Linux from the product altogether? Perhaps the hardware difference is a clever counter-move?

From day one they were including all of the drivers for XP with each unit with the expectation many/most would be reloaded after purchase.

And your evidence that this actually occurred? (eg, nmap scans of Asus Eee MAC addresses on your network.) I don't have any problem with Asus providing Xp drivers -- we've long argued that vendors should be more neutral towards the operating systems running on their hardware and that argument works both ways.

Negroponte succumbs

Posted Apr 23, 2008 23:58 UTC (Wed) by djabsolut (guest, #12799) [Link]

Either Nicholas is smoking something, or Microsoft has put its thumbscrews to use again. The main interface is Sugar, and it could be argued that it doesn't matter what is underneath. However, why pay for the "underneath" when an almost-zero cost and a considerably technically better alternative exists ?

Putting XP on the OLPC doesn't make any technical sense. Firstly, XP ("lightweight" version or not) is not exactly a good match with a laptop that is deliberately limited in resources. Secondly, and probably more importantly, the security track record of XP is like Swiss cheese -- and it's supposed to run on machines that create a wide mesh network. This will end being an effective bot network: spam, denial of service attacks, captcha cracking, etc. The potential size of this bot network can have political implications with international overtones -- president X of 3rd World Country Y might not be very impressed when he/she gets pressure from USA & EU over continuous spam originating from his/her country.

Negroponte succumbs

Posted Apr 24, 2008 4:58 UTC (Thu) by jmorris42 (guest, #2203) [Link]

> The main interface is Sugar, and it could be argued that it doesn't matter what is
underneath.

Exactly.  If you try to parse what he said it makes zero sense.  His complaint is Sugar is
immature so he is going to replace the underlying Linux with XP.  But he is totally committed
to Sugar.  Huh?

Don't have one, but don't recall any serious complaints about the underlying OS, just about
the immaturity and strangeness of Sugar.  But if you ignore the words and observe the actions
it makes alot more sense.  After so many people and entities contributed effort it would
trigger a firestorm if he announced an instant about face to Windows so he is doing it in
parts.

Sugar can't be abandoned later unless Windows gets in underneath now.  Once Windows is
underneath it will be a straightforward step to add the ability to get at the XP desktop and
then slowly deprecate Sugar.  In a year or so it will be, "Sad really, just never could get it
to work, but we have to be practical about these things.... blah blah. Remember it is all
about what is best for the children.  blah blah."

This was all planned.  There was zero chance Microsoft was going to permit millions of users
to grow up without being addicted to Windows, thus the only way this project could succeed was
with their blessing.  The trick was finding a way to make Microsoft WANT to be in so bad they
would contribute free (or almost) licenses and the porting efforts.  Lots of people were
warning about this outcome years ago.

Negroponte succumbs

Posted Apr 24, 2008 10:21 UTC (Thu) by dgm (subscriber, #49227) [Link]

I have zero evidence, but I guess that many more deals could have been signed if the XO was
Windows based. Not in the poorest countries, mind you, but in the ones that already have
computers running Windows (and educative software for Windows).

What I miss here is a bit more openness, something Negroponte should have learn from Open
Source. Is there *really* any problem with the Linux OS the XO are using? Is there *really*
any problem with Sugar? Have all those purported problems been discussed in the open?

Regarding Negroponte's selling-out, I think that he has seen Open Source as a means to an end,
that being giving the children tools for learn better, and their professors tools to teach
better. The goal has never been to promote Open Source/Free Software nor Sugar "per se", the
same way they do not promote their particular display or the mesh networking stuff. They are
cool, but are just tools.

On the other hand, all those defections of late do not speak good of Negroponte's ability to
lead the organization. But let their acts speak for them.

Negroponte succumbs

Posted Apr 24, 2008 12:41 UTC (Thu) by zooko (guest, #2589) [Link]

I suspect that to the majority of educated people in those countries (and especially the
majority of Ministers of Education and Ministers of Technology and suchlike), Linux and Free
Software for end users is understood as simply a cheap knock-off of Windows.  They probably
think of Free Software as being simply the "produced-as-cheaply-as-possible flimsy knock-off"
product, or perhaps as "the failed competitor product that is now slashing prices out of
desperation".  If they have ever heard a philosophical or strategic argument for Free Software
for end users being a better alternative than Windows, they probably thought so little of it
that they forgot about it a minute later.

So I strongly suspect that in trying to sell OLPCs to those people, the Linux factor was a
millstone.  "We don't want our children to learn how to use incompatible, failed technologies!
We want them to learn Windows like Western children do!"

Negroponte succumbs

Posted Apr 24, 2008 17:30 UTC (Thu) by allesfresser (guest, #216) [Link]

That's funny--your enquoted descriptions pretty much match my opinion of *Windows*.
Especially the flimsy part.

Negroponte succumbs

Posted Apr 24, 2008 22:25 UTC (Thu) by man_ls (guest, #15091) [Link]

To me too. Sadly we are in the minority here. Most people think that Windows is the good OS and Linux is "the bad one", as a coworker put it when I showed her my Eee. Of the rest, most think that Linux is a do-it-yourself OS not really suitable for end users. A pity, but not something that cannot change.

Negroponte succumbs

Posted Apr 26, 2008 8:27 UTC (Sat) by Cato (subscriber, #7643) [Link]

"Most people" have no opinion about Linux because they've never heard of it...  If they have
heard of it, the opinion is more likely to be "too techie" than "bad" (what does "bad" mean
for an OS anyway?).

Negroponte succumbs

Posted Apr 26, 2008 9:38 UTC (Sat) by man_ls (guest, #15091) [Link]

"Bad" means in this context "cheap", "of poor quality". I suppose part of this reputation is due to companies giving it away. Its presence in low cost computers may also be a factor. Most people don't know e.g. that the vast majority of supercomputers run Linux (although they have most certainly heard of Linux).

Negroponte on OLPC's commitment to Sugar

Posted Apr 24, 2008 13:26 UTC (Thu) by amurphy (subscriber, #29715) [Link]

This sounds like what Steve Jobs did with NeXT.

Step 1. NeXTstation - Make a really cool computer that defies the status quo.
Step 2. NeXT OS - When the computer doesn't sell, declare yourself a software company and
decouple the OS from the hardware. 
Step 3. OpenStep - When the OS doesn't sell, call it a framework and make it run on other OSs.

Step 4. Mac OS X - Sell out, reinvent and rebrand the product.

Negroponte seems to be halfway through already.

Austin

Negroponte on OLPC's commitment to Sugar

Posted Apr 27, 2008 0:45 UTC (Sun) by svkelley (guest, #37299) [Link]

I participated in the OLPC G1G1 and also purchased an ASUS Eee PC.  At the time my children
were 
three and five.  Both of them quickly gave up on the OLPC and gravitated more to the ASUS
product.  
It may be that the OLPC is targeting older children?  I don't know.  What I do know is that I
found 
the Sugar UI to be slow, very slow, and quite frankly unintuitive. I found myself trying to
figure out 
how to use it.  

Just my 2 cents.

Sean