OLPC 2.0: towards the US$75 laptop (apc)
Even before todays crop of $500-$800 mini-notebooks became the New Cool Thing, MITs vision for a low-cost Linux-powered laptop for the third world was reaching for a price point as low as US$100. It never got there, of course the product that evolved into the XO-1 sub-note sells for US$188. Even so, its notched up some 600,000 sales to date. Now MIT has its sights on a next-gen XO machine, dubbed XO-2, which it hopes will embrace new technologies while also bringing the price down to US$75 by the time the model is ready in 2010."
Posted May 22, 2008 17:47 UTC (Thu)
by i3839 (guest, #31386)
[Link] (20 responses)
Posted May 22, 2008 18:05 UTC (Thu)
by smoogen (subscriber, #97)
[Link] (19 responses)
Posted May 22, 2008 23:05 UTC (Thu)
by man_ls (guest, #15091)
[Link] (18 responses)
Posted May 23, 2008 0:41 UTC (Fri)
by smoogen (subscriber, #97)
[Link] (2 responses)
Posted May 23, 2008 6:53 UTC (Fri)
by rvfh (guest, #31018)
[Link]
Posted May 23, 2008 7:46 UTC (Fri)
by man_ls (guest, #15091)
[Link]
Posted May 23, 2008 1:12 UTC (Fri)
by dlang (guest, #313)
[Link] (14 responses)
Posted May 23, 2008 8:42 UTC (Fri)
by man_ls (guest, #15091)
[Link] (13 responses)
Posted May 23, 2008 15:32 UTC (Fri)
by khim (subscriber, #9252)
[Link] (12 responses)
Why didn't they use a distro that fits in 128 MB? I know they exist, I use them on my 128 MB laptop. Are you 100% sure? Do you have swap enabled on this laptop? What kind of browser you are using? How good it works with 128 MB? Remember: with flash swap is not an option...
Posted May 23, 2008 15:47 UTC (Fri)
by man_ls (guest, #15091)
[Link] (6 responses)
Anyway, the objective was not to run Linux in 128 MB, but to run an open system on a $100 laptop. Nobody said it was easy :D
Posted May 23, 2008 19:23 UTC (Fri)
by khim (subscriber, #9252)
[Link] (5 responses)
On the other hand, I am not sure why swap on flash is not an option. I have it enabled on my NSLU2 and it works fine! Depending on workload you can kill flash device with swap in days (worst case) or months (typical case). Of course if swap is enabled but not actually used then it's different story and the big question then is "why to have swap at all?". OLPC was designed to last years. You can just throw away your NSLU2 when it'll be dead - not an option for XO. And OLPC's biggest problem was with Firefox - it just refused to fit in 64MB or so (with 128MB of RAM you browser must use half of that because it's not the only thing running).
Posted May 23, 2008 23:49 UTC (Fri)
by zlynx (guest, #2285)
[Link] (1 responses)
Posted May 29, 2008 12:05 UTC (Thu)
by khim (subscriber, #9252)
[Link]
Possibly the fact that I paid more for that one CF card than an entire OLPC makes a difference here. If it's SLC then I'm not surprised it lasted for year. MLC is cheaper but SLC is faster AND it can support 10 times more write cycles: 10'000 vs 100'000. That's big difference. It makes no sense to use expensive flash for OLPC when you can just double the RAM for the same price thus I've assumed MLC everywheree.
Posted May 24, 2008 0:14 UTC (Sat)
by man_ls (guest, #15091)
[Link] (2 responses)
I will not throw away my NSLU2, just the USB memory (which BTW is about 5 years old and is still in working order). And then for 6 I will buy another 2 GB high speed USB memory that should work for some 25 years more. Talk about the price of consumables!
Why have swap? Because Firefox refuses to run in 64 MB RAM! No other reason: with a low footprint browser you might as well disable swap. It is a classical engineering dilemma, and one where OLPC was not creative enough -- if that is the real price increase, which I doubt because my $300 Eee comes with 512 MB of RAM.
Posted May 29, 2008 12:13 UTC (Thu)
by khim (subscriber, #9252)
[Link] (1 responses)
The links you've supplied starts with the following assumption: Wear-leveling is effective to some degree. If you are using flash for JFSS or YASS or you have some special firmware - it IS effective. With swap and no special hardware it's not. USB sticks typically have this special hardware so if you are are using not built-in flash for swap but USB stick it's somewhat different story. It is a classical engineering dilemma, and one where OLPC was not creative enough -- if that is the real price increase. It is a real price increase. We are talking about something like $2-$3 here, it's not much, but when we are talking about $100 prices $2-$3 here, then $3-$4 there and suddenly the end result is $188 laptop, not $155 laptop... And $155 is real price which "$100 laptop" meant: when the whole story started we had 1EUR == $1 USD, today 1EUR == 1.55USD...
Posted May 29, 2008 21:25 UTC (Thu)
by man_ls (guest, #15091)
[Link]
Posted May 27, 2008 9:50 UTC (Tue)
by forthy (guest, #1525)
[Link] (1 responses)
I had Linux running on a 16MB 486 machine. With X and web browser.
Yes, that's 13 years ago now, but it worked. When I updated to a Pentium
and 64MB, it even stopped swapping.
Now, I've just upgraded to a 4GB quadcore Phenom configuration to make
it stop swapping again - I had 1GB before, and it was disgusting. This is
bloat. Bloat happens, but the software from 13 years ago is still
available.
Posted May 29, 2008 12:20 UTC (Thu)
by khim (subscriber, #9252)
[Link]
Bloat happens, but the software from 13 years ago is still available. If only it were so simple. What sites can you watch with 13 years old browser? home.mcom.com? Great, but hardly useful... Sadly, but you need modern browser to use modern web...
Posted May 28, 2008 8:51 UTC (Wed)
by hingo (guest, #14792)
[Link] (2 responses)
Posted May 29, 2008 12:23 UTC (Thu)
by khim (subscriber, #9252)
[Link] (1 responses)
Nokia Internet Tablet is not OLPC. It's expensive (and it can really affect longevity of flash: SLC can handle 100'000 write cycles where MLC can only handle 10'000) and it's not designed to last ten years...
Posted Jun 1, 2008 17:01 UTC (Sun)
by shane (subscriber, #3335)
[Link]
Posted May 22, 2008 18:01 UTC (Thu)
by jhoger (guest, #33302)
[Link] (1 responses)
Posted May 23, 2008 10:58 UTC (Fri)
by NigelK (guest, #42083)
[Link]
Posted May 22, 2008 19:37 UTC (Thu)
by donbarry (guest, #10485)
[Link] (2 responses)
Posted May 23, 2008 1:31 UTC (Fri)
by drag (guest, #31333)
[Link]
Posted May 23, 2008 2:46 UTC (Fri)
by ikm (guest, #493)
[Link]
Posted May 23, 2008 2:53 UTC (Fri)
by showell (guest, #2929)
[Link] (2 responses)
Posted May 26, 2008 6:17 UTC (Mon)
by man_ls (guest, #15091)
[Link]
I wonder if the Eee and similar UMPCs would have got where they are now without the XO. My take is: probably the XO didn't make a difference there. Other devices like the wonderful Nokia Internet Tablets had nothing to do with OLPC, and there they are. Therefore Negroponte's next milestone doesn't mean too much for the industry.
Posted May 26, 2008 16:39 UTC (Mon)
by bronson (subscriber, #4806)
[Link]
Posted May 23, 2008 6:48 UTC (Fri)
by csawtell (guest, #986)
[Link] (1 responses)
Posted May 23, 2008 19:28 UTC (Fri)
by khim (subscriber, #9252)
[Link]
$60-70 were caused by quite real factors (easy and cheap oil is mostly gone by now - things like oilsands need price of crude >$50 to be profitable), but recent run to the north of $100 is probably a bubble.
Posted May 23, 2008 9:39 UTC (Fri)
by renox (guest, #23785)
[Link]
Posted May 25, 2008 14:52 UTC (Sun)
by sailor2001 (guest, #52239)
[Link]
Posted May 26, 2008 13:22 UTC (Mon)
by bignose (subscriber, #40)
[Link]
OLPC 2.0: towards the US$75 laptop (apc)
So, in short, the first version flopped.
The second version is even more ambitious. Takes a few years before it's there, but then it
will flop too.
OLPC 2.0: towards the US$75 laptop (apc)
Define flop? If OLPC were a private company it would be awash in billion dollar VC and other
funding. Getting a product out near schedule (eg not years late) and ~600,000 sales for a
startup is by any measure for that market a success. Especially when most of the people who
have done this before get squashed by the big players before 10,000 sales.
Now if you are thinking that they should have sold the 600 million or such that most CEO's
claim they will do and had Dell or HP at their knees.. then you are of course going to be
disappointed or get schadenfreude because it didn't happen.
If your objective is to make a $100 laptop and you get a $188 laptop (as the quoted paragraph says, although it actually retails for $400), then you have failed. If your objective is to create a vehicle for education and the learning materials somehow fail to materialize, you have not succeeded. Finally, if you plan to create an "open platform for learning" and you end up offering Windows on a puny machine, you fail miserably. Unit sales don't mean a thing, especially for a non-profit, don't you think?
failure criteria
failure criteria
Missing by $88.00 with a falling US dollar and other random costs is not bad. The problem is
that you can miss by $88.00 on a 2400.00 laptop and not get noticed... on a 100.00 it stands
out.
Any of the other goals that you list are things that take years to occur. Linux was not
created in one year... I figure SugarLabs and other items are further along than what Linux
was at one year later.
Unit sales mean a lot for non-profit actually. Non-profits that can't show they have some sort
of traction will not get further outside funding any more than a startup will.
I mainly say this because while OLPC has succeeded in spite of Negroponte.. calling it a
failure is demeaning to a heck of a lot of people who have worked their hearts out.
failure criteria
They should have put the objective at 100, it would show less...
I did not intend it to be demeaning. Many valuable people have already moved on, and will probably accomplish more at their new enterprises. As you say, in spite of Negroponte.
failure criteria
failure criteria
the $100 price is based on a higher volume than has been produced yet.
also the fact that they were not able to get a modern linux distro to work in the inital 128M
of ram meant that they needed to up it to 256M, things like this add up.
as for the 'retails for $400', that depends on the quantity that you are buying. when they
quoted the $100 price was was for quantities of 50,000 or more, single unit quantities are
always higher. in quantities of 10,000 they are $199.
Those are actually the reasons for the failure, not the justifications. Why didn't they use a distro that fits in 128 MB? I know they exist, I use them on my 128 MB laptop. Why didn't they reach the desired volume? $400 is still impressive, and the OLPC has broken ground in many areas, but it falls short of the expectations set by the project.
failure criteria
Which ones?
I am sure that, depending on what applications I use, it fits on 128 MB. On the other hand, I am not sure why swap on flash is not an option. I have it enabled on my NSLU2 and it works fine! Maybe the life of the flash disk will be shortened a bit more, but with modern improvements I'm not sure it is that significant.
Which ones?
OLPC was designed to last YEARS not DAYS
OLPC was designed to last YEARS not DAYS
I've been using a high-speed CF card for Linux swap for a year now and it still works quite
well. It is heavy usage because it's a Gentoo system and I compile updates on it while
running Evolution, many browser tabs and Beagle. Sometimes swap hits 600 MB and GCC works it
hard.
Possibly the fact that I paid more for that one CF card than an entire OLPC makes a difference
here.
SLC has 10 times more resources then MLC
Firefox is actually the biggest memory hog on my 128 MB system. It does not seem like the best browser for a low memory machine -- although under XFCE it eats up quite less memory than under GNOME. But on my laptop Linux runs routinely under 80 MB RAM with firefox running and GMail loaded, using a bit of swap. I don't know what the swap was used for, it didn't thrash very much.
Is 25 years good enough?
Not quite 25 years.
Not quite 25 years.
If you are using flash for JFSS or YASS or you have some special firmware - it IS effective.
Isn't this the kind of engineering problems the OLPC set out to solve?
With swap and no special hardware it's not. USB sticks typically have this special hardware so if you are are using not built-in flash for swap but USB stick it's somewhat different story.
[...]
It is a real price increase. We are talking about something like $2-$3 here
This "special hardware" cannot be that expensive, when for 6 you get a nice shiny 2 GB USB stick nowadays. Or just do it in software as you suggested above. Just spend $0.10 for every laptop and with the current unit volume you get $60,000 to pay an engineer that could develop JFSS further.
And $155 is real price which "$100 laptop" meant: when the whole story started we had 1EUR == $1 USD, today 1EUR == 1.55USD...
Why should the target be in (EUR)? If so, why didn't Negroponte announce the "100 laptop"? Most stuff sold and bought in the US have not suffered a 55% increase in price; that is why we Europeans travel there to buy a lot of stuff. And it doesn't show in inflation.
Which ones?
It's also uselss...
Which ones?
<em>Remember: with flash swap is not an option...</em>
Not true (anymore). The Nokia Internet Tablet does swapping. The trick is to combine swap data
into bigger batches, so you don't write 1000 times / second to disk, but lets say 2 times per
second. (I'm not sure of the actual numbers.)
henrik
Nokia Internet Tablet != OLPC
Nokia Internet Tablet != OLPC
I was curious, and the tablet is cheaper than the OLPC:
http://www.consumerdepot.com/products.asp?id=N800R&re...
OLPC 2.0: towards the US$75 laptop (apc)
It doesn't seem like the right moment at all to make a product announcement. So on top of
management issues the project also has serious marketing issues.
I don't agree with "the first version flopped" comment. Some 600,000 laptops are out there, we
need to make the most of it. It's not an event or a product, the laptop is part of a process
so it cannot really "flop."
OLPC 2.0: towards the US$75 laptop (apc)
I agree that this is a marketing misstep.
The best way to kill sales of a product? Announce that its successor is going to be better,
more powerful, cheaper, and not too far off.
Govts have just been given a damned good reason to hold off making that million-dollar-order
until the XO2 is ready.
OLPC had better have enough cash to tide them over until the launch of the XO2 - and that the
XO2 arrives on time, on spec, and at the price point they're targeting.
OLPC 2.0: towards the US$75 laptop (apc)
This announcement has the distinct aroma of something premature --
rushed out perhaps to blunt some of the overwhelmingly bad press
amongst the technorati for the latest cozying with Microsoft?
Negroponte has had one specialization in which he has always excelled:
creating buzz. On most issues, that buzz has had little connection
with reality. With the OLPC, the buzz actually gained traction, but
that in spite of, not because, Negroponte -- what made the difference
was a community who came together around the project, because of its
perceived (or less charitably, dishonestly portrayed) orientation.
With that community now evaporating, what will be left? Only the buzz.
Just random Media Lab types walking around with cyborg-like heads up
displays or deconstructing constructivism.
The Sugar Foundation has the best shot to retain the community -- and
other companies can quite easily team up with Jepsen et al to create
clever hardware which can run anything. What reason for Negroponte at all?
OLPC 2.0: towards the US$75 laptop (apc)
> Negroponte has had one specialization in which he has always excelled: creating buzz. On
most issues, that buzz has had little connection with reality.
Typical politician BS. Shouldn't be surprised. (I am not) It's a lot easier to get people to
follow you when you tell them what they want to hear. Politicians tell people that they'd make
the planet a perfect happy place as long as you give them the money and power necessary to
make it happen.
The OLPC is a great thing. Lots of terrific ideas and dedicated individuals did their best to
make it happen. The cheap battery technology: terrific. Making laptops safe for kids: couldn't
happen sooner. Getting a display that worked in direct sunlight. Replacing static and
expensive books with dynamic continuously updatable media. Mesh networking. ALL OF IT is just
absolutely terrific. It's the definitely the people working on the project that made that
happen.
OLPC 2.0: towards the US$75 laptop (apc)
> With that community now evaporating, what will be left?
XO-2 seems like a nice e-book device, I'd bought one or two. This version doesn't have a "kids
only" (aka ugly) design. Maybe it's not the goal anymore? Generate a lot of publicity then
start a for-profit on top of that? :)
OLPC 2.0: towards the US$75 laptop (apc)
Boy some people are hard.
Having a hairy audacious (and barely achievable) target is a good thing if you really want to
stretch the envelope. You only succeed as high as you aim. Are you saying that this would have
been a greater success if they had targeted a $200 price? I think this is a great achievement,
missing the target by only $88 is a fantastic outcome and the 600,000 sales shows that there
was a market for this product.
If Negroponte doesn't set another audacious target he is not advancing this area. If his next
device costs $99 are we going to say he failed again. This should be seen as a call to arms.
It would be great to see if the industry can meet the challenge.
Of course the project would have been more successful if the initial target had been $200! But then Negroponte wouldn't have created so much hype, attracted so many brilliant people, burnt them along the way, and so on.
Moving targets
OLPC 2.0: towards the US$75 laptop (apc)
Once power management (remember the 10hr battery life promise?), the touchpad, and Sugar are
actually useful, I'll agree that that NN missed the goal by $88.00. Until that happens,
unfortunately, OLPC appears to have missed its target by miles.
They should express the price of the thing in terms of a hard currency such as 'barrels-of-crude'. Currently it's sitting at not quite two XO-2s per barrel.
OLPC 2.0: towards the US$75 laptop (apc)
Probably one XO-2 per barrel of crude...
Remember an *education* project
Not a laptop project, so the price of the laptop is only one point..
So the question is: will the children really have a better education thanks to this laptop?
This depends
1)on the software&data available to install on the data, is-the software good enough?
Are-there e-manuals for the children?
2)on the way it is used in the long term.
I have yet to see an evaluation of (1), but I'm not convinced that the software is good
enough..
OLPC 2.0: towards the US$75 laptop (apc)
128 mb is not out of whack You don't need firefox (bloated) kazehakase is great and the idea
of olpc is for basic usage not for running flash programs.....linux dsl works great in that
computer.
OLPC 2.0: nothing to do with GNU/Linux
I see nothing in the article to indicate that the XO-2 has anything to do with GNU/Linux. Why
is this article newsworthy here?