Re: OLPC - One Laptop Per Child
OLPC_Human_Interface_Guidelines sounds good and I think some of its ideas will go into other products. But the original idea of villagers who are 100 km from mains electricity having computer access seems likely to fail.
Aubrey McIntosh wrote: Do those same villagers have cell phones?
Yes and 'village _PCs' shouldn't use more than about 6 times as much power as cell phones, that's why I think ARM is the way to go. The hand-cranking scheme seems absurd because there is no new technology that can make a mechanical watch competitive with a semiconductor one. Ie. making a cost-effective reliable suitable power generator would be more of an engineering miracle than making the rest of the OLPC project. OTOH I don't have figures for the OLPC's Goede CPU power consumption. A big colour display is also a power user ? And wifi ? As previously stated the simputer had the right idea re. hardware. Unfortunately they failed in the important 'fuzzy' but vital area: IMO aviation, radio, micro-processors [hence the computer industry] was pioneered/driven by mostly enthusiasts/hobbyists - mostly US. The collaborative [without immediate remuneration] model, which is essential for certain projects, is culturally a no-no for non-westerners. Related to the difficult-to-quantify/describe aspects, the MIT OLPC_Human_Interface_Guidelines seems to be worth a Ph.D., and reminds me of Xerox Parc. The way wifi is integrated into all aspects of operation will further boost the kids who are using it [not African villagers] and increase the digital divide. IMO if they get the Human_Interface right, the hardware will follow. The real problems are socio-political [human related], not technical. BTW in the 80's I took a HX20 [8 bit 6301 based for which I had made a P-code Pascal-subset system] and later a X86-DOS portable to Malawi and employed 6 locals as 'short hour clerical assistants' to work on computing projects. I had the last electrified cottage on the east side of the lake. One of the biggest problems was getting scribbling paper, for the 9 months I spent there. Chris Burrows wrote:
Oberon will have its time and place but perhaps not exactly in its present form. Maybe a universal limited-resource device is the sort of launching pad that will make a significant difference.
Technology is moving so fast that limited-resource is becoming insignificant ? What remains are the human determined factors: we've still got 10 fingers, English [26 leters ++] is the dominant text, it's much easier to recognise than remember...etc. Fortunately much of Wirth's ideas are based on these universal human attributes rather than on transient technology. With my African ISP chaos, I don't know if I've already posted the text below ? =============== To: oberon@lists.inf.ethz.ch,crglur@gmail.com Subject: Re: OLPC - One Laptop Per Child Aubrey.McIntosh wrote:
There was the NIC computer that used Linux and booted from a CD, and there was no "disk". It did have ethernet. I booted BlueBottle onto that machine. They appear to be defunct.
I also have a LEX CV860A machine on my desk and it runs BlueBottle. I love this machine. This machine is available for about $250, and has no mechanical parts (heat sink, no cpu fan, cflash, USB memory stick)
I thought there would be some advantage to the Oberon community to focus on this class of machines, and I even corresponded some directly with some of the BlueBottle group.
I guess I feel jaded after maybe 15 years of wondering how to make Oberon more accepted. What was your sense of something new or compelling about the OLPC machine?
As MIT-OLPC said they are not in the computer making business, but in education. It's about the 'digital divide': until recently smallpox was a problem - AFAIK mainly USA fixed it; they could have fixed polio too, except for the Nigerian Islamists saying the innoculation plan was a racist scheme to sterilize their females. Then there's bird flu which travels globally and is going to kill you. Since the 1st world can't remain immune from eg. Africa and it's been proven that the Marshal Plan doesn't doesn't work with them [Haiti has been 'independant' for over 200 years], so G7 want to try magic technology to uplift/educate the 3rd world. Importantly, health/medicine is now seen increasingly to depend on knowledge and self managment, rather than drugs as it was before the '60s exposure of the dangers of smoking. The extreme example being HIV/AIDS. Undoubtedly computers & the net facilitates the cheap spread of knowledge. Hence the OLPC. It was planned/advertised to be supplied for US$100 - ordered 1 million minimum quantity. Now it's said that Lybia bought a packet for US$135 each. Some reports tell of costs escalating to US$150. Of course with India & China's population, a million is not much. But then you must have non-mains-electric operating. OLPC don't seem to give current consumption figures for their 'goede CPU'. But AFAIK the low power capable CPU is the ARM. Which is what the simputer chose. And which is apparently also mature for linux. Obviously the whole project centers on linux. The simputer designers had a nice idea: each user would have his own 'flash card' [the flash-mem technology had not yet exploded at that time]. By that method each scholar/user could have his own data for minimum cost, and they could time share. Like buying a bus ticket when you need it instead of owning a car. Chris Burrows wrote:
Oberon will have its time and place but perhaps not exactly in its present form. Maybe a universal limited-resource device is the sort of launching pad that will make a significant difference.
I could never see why Oberon should be more 'limited-resource' than eg. C. Probably the non-preemptive OS ? Isn't BB/Aos preemptive ? OLPC apparently describes a round-robin scheme too ? I used to think N-O's round-robin scheme was fine, but for inet, I see it's painfull: when I call news and it takes 60 secs. to reply, I've got to pay telco costs while a have useless waits. By contrast with linux, I can just send off for 10 http pages via a lynx script, while I'm fumbling manually with mail/news. == Chris Glur
I could never see why Oberon should be more 'limited-resource' than eg. C.
No - but C hasn't got the benefits of 15 years of additional programming language experience and knowledge that influenced Oberon's design. It was bad enough that programmers in the 70's and 80's were subjected to the trials and tribulations of C - surely we can do better than that in 2006? C# is a notable improvement but still has inherited some of the deficiencies of C. A good example of the use of Oberon (admittedly a subset) in a limited resource environment is Wirth's model helicopter work. http://www.uav.ethz.ch/research/projects/closed/project_OLGA Chris Burrows CFB Software http://www.cfbsoftware.com/gpcp
No - but C hasn't got the benefits of 15 years of additional programming language experience and knowledge that influenced Oberon's design. It was bad enough that programmers in the 70's and 80's were subjected to the trials and tribulations of C - surely we can do better than that in 2006? C# is a notable improvement but still has inherited some of the deficiencies of C.
Check out Zonnon http://www.zonnon.ethz.ch it is an even better selection.
--- easlab@absamail.co.za wrote:
Yes and 'village _PCs' shouldn't use more than about 6 times as much power as cell phones, that's why I think ARM is the way to go. The hand-cranking scheme seems absurd because there is no new technology that can make a mechanical watch competitive with a semiconductor one. Ie. making a cost-effective reliable suitable power generator would be more of an engineering miracle than making the rest of the OLPC project.
How about a hamster power generator? :) Here's a kid who designed one for a cell phone. http://www.ananova.com/news/story/sm_1513938.html I agree with the assertion that an ARM would be a better chip. As for handcranks I suppose the big question is how many minutes of use to you get for one minute of cranking? Also maybe a "foot pedal" might be better. Solar charger would be optimal, but that would almost double the price.
OTOH I don't have figures for the OLPC's Goede CPU power consumption. A big colour display is also a power user ? And wifi ? As previously stated the simputer had the right idea re. hardware.
Unfortunately they failed in the important 'fuzzy' but vital area: IMO aviation, radio, micro-processors [hence the computer industry] was pioneered/driven by mostly enthusiasts/hobbyists - mostly US.
The collaborative [without immediate remuneration] model, which is essential for certain projects, is culturally a no-no for non-westerners.
Ok. You've brought up the failure of the Simputer a couple of times now. I'm sure you've followed this closer than I have. But I'm not sure how your categorizing the failure? From what I've seen they managed to get a completed product though they didn't get the sales they wanted. So when you talk about the lack of collaboration are you talking about in development of the Simputer itself...or lack of third party development...or something else?
Technology is moving so fast that limited-resource is becoming insignificant ?
Well the "power/performance" curve is always there. As the saying goes "Intel giveth...MicroSoft taketh away." Check out Vista. It adds some features like semi-transparent Windows (something BlueBottle has had for some time now as well as some other alternative OS's) but what a cost! 1 GB of RAM and 256 MB for the video card just to comfortably run the OS? Of course the other touted feature of Vista, better security, may be worth the upgrade. Oberon, to my knowledge currently lacks a security model. I think one could have been built back when the object model was "slim binaries" because that model could differentiate between "safe" and "unsafe" modules. And there are IMO a lot of people who might be interested in a "volksputer". Something that's going to run decently, let them do some internet surfing and home office work, maybe some educational stuff but not cost more than a cheap DVD player. But I think the keys are managed expectations and system responsiveness. It doesn't matter have fast a computer is. What matters most is how fast it SEEMS. For instance I love Oberon's almost instant bootup.
As MIT-OLPC said they are not in the computer making business, but in education. It's about the 'digital divide': until recently smallpox was a problem - AFAIK mainly USA fixed it; they could have fixed polio too, except for the Nigerian Islamists saying the innoculation plan was a racist scheme to sterilize their females. Then there's bird flu which travels globally and is going to kill you.
Side note, I'm not at all worried about "bird flu". Studies have already shown that the Korean dish "kimchi" may be effective against it. That and garlic which is a strong anti viral. http://news.bbc.co.uk/2/hi/asia-pacific/4347443.stm As for vaccines, I quit trusting vaccine makers after U.S. Senator Bill Frist tried to set up a vaccine agency that would be excempt from the freedom of information act. Note that even the CIA and FBI aren't excempt. They also have mercury, a deadly poison, in vaccines for children. Some say that causes autism. I don't know. But autism rates are exploding in this country and nobody knows why.
Undoubtedly computers & the net facilitates the cheap spread of knowledge. Hence the OLPC. It was planned/advertised to be supplied for US$100 - ordered 1 million minimum quantity. Now it's said that Lybia bought a packet for US$135 each. Some reports tell of costs escalating to US$150.
So the OLPCs are already being sold? I didn't know the project was at that stage yet.
Of course with India & China's population, a million is not much. But then you must have non-mains-electric operating. OLPC don't seem to give current consumption figures for their 'goede CPU'. But AFAIK the low power capable CPU is the ARM. Which is what the simputer chose. And which is apparently also mature for linux. Obviously the whole project centers on linux.
The simputer designers had a nice idea: each user would have his own 'flash card' [the flash-mem technology had not yet exploded at that time]. By that method each scholar/user could have his own data for minimum cost, and they could time share. Like buying a bus ticket when you need it instead of owning a car.
Oberon will have its time and place but perhaps not exactly in its present form. Maybe a universal limited-resource device is
Chris Burrows wrote: the sort of launching pad
that will make a significant difference.
I could never see why Oberon should be more 'limited-resource' than eg. C. Probably the non-preemptive OS ? Isn't BB/Aos preemptive ?
Yes.
OLPC apparently describes a round-robin scheme too ? I used to think N-O's round-robin scheme was fine, but for inet, I see it's painfull: when I call news and it takes 60 secs. to reply, I've got to pay telco costs while a have useless waits. By contrast with linux, I can just send off for 10 http pages via a lynx script, while I'm fumbling manually with mail/news.
Native Oberon doesn't HAVE to be that way. Oberon V4, which also uses cooperative multitasking, can do background web fetches. You can keep doing whatever you want while a page is being downloaded and rendered. It takes a bit of thinking to turn a normal procedure into a task that can be run in the background with single process multitasking but it can be done. On the flipside preemptive multitasking can be done on resource constrained machines. The Amiga, for example had preemptive multitasking back in the 1980s. Regards, John M. Drake ____________________________________________________________________________________ Do you Yahoo!? Everyone is raving about the all-new Yahoo! Mail beta. http://new.mail.yahoo.com
participants (4)
-
Alan Freed -
Chris Burrows -
easlabīŧ absamail.co.za -
John Drake