Radio and broadcast
Broadcast texts over shortwave radio
A major challenge is providing access to information through the green machine. Access to the Internet in most of the regions of the world where children will benefit from the OLPC program is usually limited. Here's a hardware+software idea that could make access to information truly ubiquitous. It goes for _access_ as opposed to speed. The idea works synergistically with other ideas such as Jason's idea for p2p distribution of texts and mesh network access to the Internet.
Rationale: I see the green machine as (1) a tool for accessing information and (2) a tool for self-driven learning through investigation. A problem with the former role is "from whence cometh the information ?". With the limited storage resources, "pre-loading" the information on the machines seems like an untenable approach. The crux of this proposal is that having easy and reliable access to a small library of information will be more useful than having spotty access to the entire Internet, even through mesh networking. The idea focuses on access to information, rather than access to the Internet, the latter and Information not always being synonymous.
Idea: Broadcast texts over shortwave radio. Have the texts broken up into chunks, and all the texts from a library interleaved in time. If possible, transmit different interleaving sequences on a set of different channels. The basic circuitry for a shortwave radio demodulator can be very simple, particularly if the decoding is done in software sitting behind a straightforward RF and analog front-end. When a child wants information on a topic, the implemented system captures fragments of a text relating to the topic off the air, and over time, the whole text is collected.
Example: Purely as an example, let the library of texts being transmitted consist of (1) Project Guttenberg texts (2) (one thing I wish I had access to as a kid:) mathworld.wolfram.com (3) Some free encyclopedia. All the content is broken up into packets of some size, encoded, all the packets interleaved according to some scheme, and the entire interleaved stream repeatedly transmitted. The infrastructure for transmitting the data streams need not even be in the nation or on the continent where the recipients are. Let, for example, the entire stream take two months to transmit. 4GB of encoded data at 1200 baud will take ~40 days. 1200 baud for a packet radio seems feasible, though I can't say how complex the hardware would be. Bear in mind we're only talking about receiving, not a full-blown modem.
Extensions: You can imagine having N different interleavings, which are transmitted on N different channels. When a child wants information on some topic, the request is sent out to at least N other green machines, over the mesh wireless network. All the machines start picking up chunks for the text on one of the N different channels. The chunks are sent to the requester over the mesh network.
That's the top level idea. I think it is feasible to implement with minimal hardware,
and the algorithms and their software implementations to go behind the hardware are interesting issues, many of which are already well understood or solved.
--Pstanley 10:52, 1 February 2006 (EST)
Existing technology that can be adapted
There are various existing technologies that could be adopted or adapted to achieve this.
Digital Signal Processing for Ham Radio
The Ham Radio community has a vast amount of experience with weak-signal transmission (and reception of data) beginning with slowscan TV back in the 60's. Phil Karn, the developer of a popular DOS application to run TCP/IP over shortwave, has done some work that could be used. The Geode chipset does have MMX instructions which can be used to do DSP functions.
HF radio
The ham radio community has mature technology for sending data over long distances at low speeds over narrow high-frequency (3-30 MHz) channels using PC sound cards as data modems. This technology could be used for e-mail to remote regions.
I think this would be a good idea since some children live a mile or more in there towns and villages. Having data travel over long distances will help them keep in touch with each other through the network. I'm sure thought this will have to be approve by the IEEE as a new networking standard, but I'm sure this can be done.
USB addons
Well, first of all, the OLPC has 4 USB ports. There is no reason why someone could not manufacture such a radio system as an accessory for the OLPC, to be used in those countries and those regions where this would be a good idea. This would most definitely be a dumb idea in the capital city of Thailand, for instance. But in mountainous western China it would be a very good idea, indeed.
Vertical retrace data channel
One way to mass download info, at least in areas that have TV reception, is to use the vertical retrace interval to send data.
Miscellaneous
Support for this and mention of telesoftware and of MHP
Yes, good idea.
You might like to consider using telesoftware, that is the unidirectional cyclic broadcasting of software and its selective use, to broadcast software packages for the laptop as well.
The following webspace about the DVB-MHP (Digital Video Broadcasting - Multimedia Home Platform) interactive televison system might be of interest.
The MHP system broadcasts Java programs to interactive television sets.
The following idea, not implemented as far as I know, might be of interest in relation to digital radio.
http://www.users.globalnet.co.uk/~ngo/tor00000.htm
William Overington
5 March 2006
There is now a page about the possibility of an MHP to laptop interface
There is now a page about the possibility of an MHP to laptop interface.
[[1]]
One way connection to internet helps nothing. This is passive usage like watching tv. Software distribution with USB is easy.
- Yet in areas of the world where a two-way link is impossible, a one-way system where lots of software and data such as electronic textbooks can be received from direct to school broadcasts from a geosynchronous television broadcasting satellite and stored localy, telesoftware could be very advantageous.
Historical note about the ATS-6 experimental satellite
Some interesting experiments in using direct satellite broadcasting for education took place in the 1970s using the ATS-6 experimental satellite.
http://www.nasm.si.edu/research/dsh/artifacts/CS-ATS6.htm
William Overington
6 March 2006
Distance and the Mesh Network
As I understand these laptops will have intergraded wireless with the 802.11 b and g standards. The problem I see is that in some countries, like Africa, the children live sometimes a half mile or so apart from each other. The 802.11 b and g IEEE standard only has a reach of 100 feet or more with out modifications. This means the only time the children will be networked with each other is when they are at school. What needs to be done is the wireless in each laptop needs to reach up to at least a mile so the children’s computers are always connected to the network even in their home. The wireless b and g standard can be mod up to 10 miles of reach but to do that requires either a large antenna or an amplifier that would consume too much power. My suggestion is that you create a new wireless networking standard with low radio frequencies like every one has been suggesting. I would like to see it so the children’s computers are always part of the network when they are at home, because as I see it most of the programs on the laptop except the office programs can’t work with out a network. --68.66.231.5 06:52, 31 May 2006 (EDT)
Speeds and data integrity
If I remember my modem hardware correctly, 1200baud is simple FM over half of a telephone channel. This would be more like a FAX channel where the entire frequency width of the channel is available; all normal fax modulations have a baud rate of 2400, with bps of 4800 to 14400.
The processor in this machine would be more than enough to demodulate this in software and decode some serious forward error correction eg par files
62.252.0.11 17:02, 17 March 2006 (EST)