One of those believe it-or-not stories. News.com writes:

The technology is based on the interaction between liquids and solids on a very small scale. The surface of the solid receives a small charge that attracts opposite-charged ions in the liquid and repels like-charged ions. The process creates an electric double layer (EDL)--a thin liquid layer with a net charge that ranges from a thickness of several nanometers to a few micrometers.

Professors Daniel Kwok and Larry Kostiuk from the University of Alberta created channels similar in size to the EDL and forced liquid through the channels, resulting in a movement of net charges downstream. Because the ions that are repelled by the surface move faster than the ions that are attracted to the surface, a current is generated, which leads to a voltage difference across the ends of the channel if the solid is a nonconducting material.

The power generated from a single channel is extremely small--a 30cm column of water will produce one to two microamps (one millionth of an amp)--but the researchers envisage millions of parallel channels being used to increase the power output to a level sufficient to power small electrical devices such as mobile phones and calculators.

"The applications in electronics and microelectronic devices are very exciting," Kostiuk said. "This technology could provide a new power source for devices such as mobile phones or calculators which could be charged up by pumping water to high pressure. What we have achieved so far is to show that electrical power can be directly generated from flowing liquids in microchannels."

David Galbraith writes:

Phil Wainewright has a very thorough and thoughtful piece on Loosely Coupled. Phil looks at Webmethods' recent acquisitions within the broader context of what happens when enterprise software all starts to communicate through standardized web services based API's, rendering a whole industry (EAI) obsolete.

Its an interesting mirror of what is happening at a grass roots level with syndicated content and standardized API's for weblog style publishing systems. RSS aggregators beware - the future is not in aggregating or aggregation tools, but services building value-add on top.

Atanu Dey points out the fallacy of the phrase "digital divide":

It is not the digital divide that is preventing the poor from benefiting from ICT. It is the fact that they are poor that is preventing them from benefitting from ICT. Not just benefitting from the use of ICT, the poor also are not benefitting from the advances in medical technology, in cosmetic surgery, in plasma tv technology, ad nauseum. It is not the digital divide, stupid, it is an income divide, it is a wealth divide, it is an opportunity divide.

If the poor had money, they would not be poor, and like all non-poor, would be able to buy all sorts of stuff -- including, but not limited to -- digital gizmos. They would buy education, clothes, food, houses, cell phones, cd players, dvd players, plasma tvs, and computers. There would not be a digital divide. It bears repeating: the digital divide is not the cause of poverty nor is it the cause of the persistence of poverty. The digital divide is a result -- an effect, a consequence -- of poverty.

Now coming to India: India does not have a digital divide.

If a vanishingly small number of people have something, there is no divide. For instance, it is pointless to talk about a BMW divide: we are all in the same boat when it comes to having BMWs and therefore there is no divide. So also, to a first approximation, Indians don't have access to the Internet, except for a few million people. And the few million who do have it, have to pay inordinate amounts of money to get a slow uncertain connection.

Dan Gillmor eloquently summarises the change that is coming to the world of cellphones: "The key feature of these new phones is how they become almost a hybrid of the two major communications devices of the late 20th century, voice phones and personal computers. They're phones plus computing platforms, taking advantage of specialized network services, but not so complex as devices that they become as unwieldy as PCs or so simple that they can't be adapted to other uses."

For Trujillo, the CEO of Orange, the mantra is simplicity. It has to be easy to use, intuitive and so compelling that customers will spend more time online -- on today's and tomorrow's digital mobile networks. A wide range of applications will be available, but customers will choose a relatively small set that are specific to their own needs.

No longer will mobile carriers have to create large market segments that put people with differing needs into a single category, he said in an interview. Each customer will be "a segment of one.''

What makes this vision of the future different is the insistence that the customer actually gets to pick and choose.

Joi Ito "maps the space" we are all part of. "The x axis is the "context". IE low context is stuff like CD's and books which don't change, are worth approximately the same amount to most people and don't have much timing or personal context. The far right is very personal, very timing sensitive, high context information such as information about your current "state". Then there is everything in between. The top layer is the type of content sorted by how much context they involve. The next layer is how they are aggregated and syndicated. Below that are substrates that are currently segmented vertically, but could be unified horizontally with open standards."

This is something which requires "deep thought".

WSJ fastforwards to 2004, and suggests five things that we'll do next year:

1. You'll subscribe to or buy from a legal online music service.

2. You'll have a conversation with your roommate/significant other/etc. about scrapping your landline in favor of your wireless phone.

3. You'll set up a home wireless network.

4. You'll own something with GPS in it.

5. You'll routinely buy something online that you'd think was crazy to buy that way today.

WSJ writes about the culture spawned by the call centres in Philippines - something we will probably see replicated in India.

There are 30,000 people answering phones and e-mail queries in Manila, doing work -- for $600 to $800 a month -- that generally pays better than bookkeeping in a bank or similar white-collar employment. That figure could double over the next couple of years, call-center operators predict, creating a middle class that actually stays in the Philippines rather than one that must emigrate to work overseas, as millions of Filipinos have done.

At 2 a.m. on a recent Saturday, Sherylyn de la Santos bit down on a microwaved cheesedog on the steps of a 7-Eleven store. Her friend Gabriella Manalo held her Coke. It would be a midafternoon break in Baltimore, but in Manila it was time to hit the booming all-night cornershops for some refreshment.

So far, there is no shortage of people who want to work on U.S. time -- companies such as eTelecare and People Support get as many as 200 applications a day, thanks in part to an unemployment rate of nearly 13%. But those who get jobs often find it difficult adapting to "the bubble."

Che Che Montero has been working in a call center for nearly a year. "Most of the time I feel isolated," she says. "You can't tell your old friends what kind of day you've had because they are either sleeping or at work."

Staying awake is less of a problem, the 25-year-old physical-therapy graduate says. "I don't get sleepy at the office. There's a lot of irate callers to keep me awake at night."

For others, there are things to do while everybody else is going to work.

Indian Express lists the Indian nominations for awards at the World Summit on Information Society in Geneva from December 10-15 later this year. The winners will be announced shortly.

CATEGORY: E-learning NOMINEE: Totaliteracy (Tata Consultancy Service)
CATEGORY: E-Governance NOMINEE: E-seva project in Andhra Pradesh
CATEGORY: E-Biz NOMINEE: ITC’s kiosk-based E-chaupals
CATEGORY: E-Culture NOMINEE: NIIT’s Hole-in-the-Wall
CATEGORY: E-Science NOMINEE: Simputer (IISc, Bangalore and Encore Software)
CATEGORY: E-Health NOMINEE: TeleDoc programme (Jiva Ayurveda)
CATEGORY: E-Entertainment NOMINEE: Contest2win
CATEGORY: E-Inclusion NOMINEE: N-logue (IIT Chennai)

The thin clients shift the onus of processing and storage to the server. Luckily, the relentless advances in computing driven by Moore’s Law means that we can leverage this dramatic increase in computational capabilities on the server by using the newest desktops as servers in enterprises. While Intel and AMD have distinct desktop and server brands, in the context of small- and medium-sized enterprises (SMEs), that distinction is not as meaningful any more.

The thick server which resides on the enterprise network should be capable of handling all the computing activities. There are two ways to size the server, given the size of the organisation: one is to go in for a single server which has the higher-end CPUs (like the Intel Xeons), while the other is to scale horizontally and go in for multiple desktops in a server cluster.

Many of the smaller companies would only need a single server, so it does not matter. But for the mid-sized companies, the second choice may be a better approach, because if offers greater reliability (no single point of failure – which can be a justified criticism for a server-centric computing solution) and scalability (just add another computer as the user needs increase).

For the local networking, there are two possibilities: the use of WiFi within the organization or doing cabling. WiFi would be a preferred option because it reduces the installation time – cabling can be a cumbersome process.

The software on the server comprises of four layers: the operating system (typically Linux), a distributed file system (in the case of multiple computers working together), terminal services (to provide support for the thin clients) and the applications. All the software needed to make the thick server a reality exists in open-source today. What has not happened is the complete and seamless integration of these applications to make for an integrated system.

The cost of such a thick server would vary from about Rs 25,000 (USD 550) for the minimum configuration for a small organisation (of 5-7 users), and go up to Rs 150,000 (USD 3,300) for a 50-person organisation. A good assumption is that it will cost Rs 5,000 (USD 110) per user in server costs for a small business, with the cost going down to Rs 3,000 (USD 65) per user for larger enterprises.

The two big advantages of this thin client-thick server architecture are that the clients never ever need to be upgraded, and management of the IT setup boils down to management of the servers only. If the network connectivity is present, the servers can in fact be managed remotely, eliminating the need for trained engineers to be located onsite, which is always an expensive proposition from the SME viewpoint. This architecture also benefits from the continuing commoditisation of IT hardware – it should be possible to get at least 50% greater processing power and storage a year down the line for the same investment.

Thus, it is possible to put together a complete hardware solution (desktop thin client and server) which costs no more than Rs 10,000 per person in upfront investments. This is a 60-75% reduction from today’s costs, and forms the first foundation for the SME IT architecture.

Tomorrow: Systems Software Architecture