Paw Prints: Writings of the maddog

By now you probably have heard about the “pad” computer designed in India that is being touted as costing thirty-five U.S. Dollars to manufacture. While there is very little in the way of technical details about it, some information has been published that says it consists of:

• An ARM9 Architecture Processor from Freescale (I.MX233): 5 USD

   

• Memory: 3 USD

   

• WiFi b/g: 4 USD

   

• Other “discrete” components: 3 USD

   

• Battery: 5 USD

   

• 7” 800x480 resistive touch screen: 15 USD

   

for a total bill of materials: 35 USD, and rumors that in the future this will drop to 20 USD and even 10 USD. The system is “Linux based”, but does not say if it is based on Android, ChromeOS or some other Linux-based distribution. There is also no mention of persistent storage other than the fact it has a USB connection that could be used for flash.

As other reporters and blogs have pointed out, there is no mention of the PCB for the motherboard (assuming it has one), nor the assembly, packaging, transportation, testing, returns, etc. The reports run hot and cold about how much this system will really cost, whether it will have enough power in the system to be useful, whether it will ever really be produced, whether it is “rugged enough”. There was no mention of the operating temperatures of the unit. There are a lot of very hot places in India, and of course the system has no fan. There were also lots of comparisons with the OLPC, which I feel are unfair since the OLPC was solving slightly different problems, and was a leader in the effort.

I am optimistic about the Indian “Pad” (hmmm, maybe they should call it the “IPad”) for several reasons. In 1969 I saw my first electronic hand-held calculator. It had four functions: add, subtract, multiply and divide. It may have had a square-root key, I can not remember. I do know it used six “D-cell” sized batteries. had Nixie tube readouts and cost 400 USD. In 1969 that was a LOT of money, way more than the average college student could afford, so most of us used slide rules (themselves often costing more than 100 dollars for three pieces of wood layered with plastic that slid back and forth).

A few years later calculators with LED numerals, that could do all of the previously mentioned functions plus have a memory, percent key, and ran off AA size batteries were available as gifts from a bank when you opened up an account.

A few years later the same “four-banger” style calculators, now running off solar power, were given away as presents at trade shows...and shortly after that they appeared in the “convenience line” in supermarkets.....something you might buy on the spur of the moment, much like candy.

Of course there were more expensive, complete calculators as there are even today, with Hewlett Packard's RPN calculators probably being the most famous. A quick search found that a Hewlett Packard's calculator was selling new for a range of 38 USD to 60 USD from different outlets on the Internet, while another model was selling from a low of 8 dollars to a high of 24 dollars.

I applaud the Indian effort, just as I applauded the effort by OLPC to create an ever-lower priced entry into what people hope is a more inexpensive way and effective way to deliver information to the masses of people unable to afford a 400 USD laptop or 300 USD net-book. In some countries the cost of these devices is almost doubled through import duties levied by the government, so every dollar in cost savings is effectively two dollars saved by the end customer. And for people who make a very low wage (or who are unemployed), this amount effectively doubles or quadruples again. Having the “pad” manufactured in India will reduce those duties to zero, and the pride of having developed something in their home land will encourage younger Indians to contribute to computer engineering.

I am undaunted by the people who point out that the “real manufacturing cost” may be higher. Even if the final cost is double the 35 USD, it still remains much more affordable than other solutions, and in time those parts will become even less expensive. Will it ever reach ten USD? Perhaps not, but even at a solid 35 USD it may be “cheap enough”.

The real issue that I see has not been mentioned in other places, which is the assumption that wireless networking will not only be very low cost, but of sufficient bandwidth and presence everywhere to allow heavy use in dense urban environments. Already we have experienced 3G congestion in places like New York City where people are downloading lots of data from the Internet to their phones and notebooks. Who is going to supply (and control) the wireless transmitters and wireless network that will be the backbone of these tablets. I have seen no mention of this in the articles. Wireless mesh? Without a significant number of back-hauls the amount of computational and network power needed to maintain the routing tables for the very mobile tablets creating the mesh might overwhelm the rather low-power processor.

People point to the “cloud” and wave their hands a lot when they talk about the backup computer power and storage for pads like these. “Build the pads, and the servers will come” to parody a phrase from the movie “Field of Dreams”.

I traveled to Uruguay when they were deploying OLPC laptops at the rate of 1600 per day. They were building their own servers, designing their own antennas, creating their own access points and otherwise working themselves to death to deploy a much smaller number of OLPCs to a much less dense area. And the OLPCs did have persistent data storage, so the children did not have to store everything across the net. Nevertheless, with a lot of hard work the laptops were completely deployed.

Will I ever buy an IndianPad? Probably not, since my needs tend to run to something a bit more powerful, but I am glad to see them fund the effort. I probably will never buy an Apple iPad either, but for different reasons. Perhaps some of their technologies will find their way into a future notebook or cell phone I may buy.

I am not going to disparage India's effort, however. The Indian people have thrown a glove in challenge just like OLPC did. Instead of lining up in the gauntlet to whip and beat them we should be cheering them on and helping them to solve these problems. A project I am working on, Project Cauã, may offer some solutions for private financing for the mesh bandwidth needed. Some of these solutions for the Indian “pad” would probably help other efforts, such as OLPC.

Did we ever build colonies on the moon? No, but the efforts to get to the moon drove solutions that benefit us all. This should not be viewed as a lose-lose game, but a win-win game.

Comments