Would you pay $400 for a handful of microchips and, armed with only a
circuit diagram, build your own cellphone?

With elegant, powerful
phones already on sale for a fraction of the price it’s not something
that will appeal to many.

Yet despite the cost and inconvenience, a growing group of techies are
attempting to build the first practical home-made cellphones. They hope
to spark greater innovation in cellphone design and, more crucially, in
the software that makes the phones work. Their aim is to develop a
critical mass of free software that will lead to a flowering of new
cellphone applications. Some foresee phones acting as affordable
hand-held computers running novel applications tailored to the needs of
the developing world.

The movement is riding on the back of a burgeoning market in wireless
devices for machine-to-machine communication. These devices are, in
effect, stripped-down cellphones, and a typical application combines
them with position sensing systems in trucks, which can then be tracked
while on the move (New Scientist

, 20 December 2003, p 30). To turn one of these basic cellphone modules
into a fully functional phone suitable for chatting to friends, you add
a microprocessor, typically running the open-source Linux operating
system, along with a speaker and microphone, screen, battery and keypad.

"Five years ago, you wouldn’t have been able to build a cellphone,"
says Surj Patel, a telecommunications engineer who now spends his spare
time building cellphones. "You could have bought bits and pieces, but
you would have needed a million-dollar lab to put everything together."
Even now, it’s not easy. Designing the circuits and writing the
software is a slog. Moreover, home-brewed phones are clunky, oversized
and burn through a battery in half an hour.

So what’s the motive? For many, it is the urge to play around with
devices that are an integral part of all our lives. They would like to
add modules such as radio frequency identification (RFID) readers to
scan the increasingly widespread wireless ID chips, or GPS units so
that a phone knows exactly where it is. They can then program custom
capabilities into the devices: to allow them to spot which of their
friends are nearby, for example.

Such tasks are beyond the capability of most ready-built phones. RFID
readers on phones are still rare, and though many existing phones have
GPS units they can often only be used by paying a monthly fee to the
network company. The cost of writing new applications is another factor
favouring home-made phones. Since most cellphone makers keep the
software that controls them a closely guarded secret, high licensing
fees and royalties effectively bar everyone except large, well-funded
companies from modifying the hardware or coding their own applications,
Patel says.

Deva Seetharam, an engineer at sensor company TagSense in Cambridge,
Massachusetts, ran into this problem while developing RFID readers for
commercial cellphones. "There is no freedom for users, researchers and
hackers to build anything," Seetharam says. "So I said, OK, I’ll build
something so people can customise the phones the way they want."
Seetharam teamed up with Patel to build their own handset, which they
named TuxPhone after Tux, Linux’s penguin mascot.

Most of the enthusiasts working on their own phones are committed to
the open source ethos that has driven the development of Linux and many
of its applications. The principle behind open source software is that
anyone is free to use or modify the software code, on condition that
their modified code must in turn be freely available for others to
modify if they wish. In this spirit, Patel intends to make public the
schematics for his phone once he has the details ironed out.

Vibrant communitySimilarly, Cina Hazegh, a graduate student at the
University of California, Irvine, and Nathan Seidel, who owns the
electronics supply company Spark Fun Electronics in Boulder, Colorado,
have teamed up to create a website called opencircuits.org for sharing
wiring diagrams.

The pioneers are convinced that if cellphone software becomes more
open, it will foster a community of programmers who will create new
applications and software add-ons. An example of how powerful this
approach can be is the open source web browser Firefox, which has long
had a slew of add-ons available for free, most of them coded by
individuals to add specific features not included in the base program.
These include advert blockers or add-ons that meld hits from several
search engines into a single page. In contrast, Microsoft’s Internet
Explorer and Apple’s Safari have little tweakability and less vibrant
communities of programmers working on extensions.

Patel has created a cellphone application called Ringfo that can be
used to call a toll-free number and enter a book’s ISBN or a CD’s code
(http://ringfo.com

). This calls up a computerised voice reading key snippets of
information from online retailer Amazon’s website, such as a book’s
average rating and prices for new and used copies. Because this voice
call does not incur data download charges, Ringfo is cheaper and easier
to use than a cellphone’s web browser, Patel says.

Ringfo is just the start. With an open source phone, Patel says he
could experiment with more efficient interfaces: for example, one that
allows the user to issue a command by voice and receive a response on
the screen. Or the phone could look for book reviews on a social
networking system such as myspace.

Another possibility is to enhance cellphone address books so that they
work in a similar way to instant messaging systems. The phone could
then tell you, for example, when a friend’s phone is switched off or
busy. "These applications are not developed because there’s no
inexpensive open platform," Patel says.

Still another approach could be to exploit the powerful microprocessors
inside even modest cellphones to run software tailored to the needs of
people who can’t afford an ordinary PC. "In the developing world,
mobile phones are a really promising computing platform," says Tapan
Parikh, a computer science student at the University of Washington in
Seattle.

Phones have advantages over PCs, he says: "They are cheaper, they are
smaller, they have a battery, people are used to using them, and they
give this immediate connectivity and voice feature." Parikh is testing
a system in rural India that uses camera phones to record bookkeeping
details for the microfinance groups through which local people pool
their money and distribute loans. His system uses Nokia phones that run
the Symbian operating system, and though Symbian is fairly accessible
to software developers it is not ideal. If cellphone software were more
open, he says, there would be more potential to explore other computing
applications.

"It’s really important to get a developer community together and have
some openness in experimentation," Parikh says. That way, local
developers stand a better chance of meeting the needs of local people.
"There’s no way for a centralised application developer to have an idea
of what applications would be useful in rural India."

Dump that operating systemMason Inman As an alternative to building
their own handsets to experiment with new cellphone applications,
engineers are attempting to bypass the existing operating system on
phones and install a new one. It’s not just independent-minded
programmers who welcome this approach. The phone manufacturers would
benefit too, as they would no longer have to pay operating systems
licence fees.

Though a few engineers have already installed Linux on commercial PDAs
and smartphones, it takes a great deal of effort to make the software
work, and even then it cannot always control all the phone’s functions.
"It is really a stopgap until we can see a device like the TuxPhone
make it big," says Matthew Mastracci, a software engineer who was the
first to run Linux on a Treo 650 phone. He and others are working to
install the software on as many phone models as possible, to prove it
can work. He hopes it will demonstrate to manufacturers that Linux is a
viable alternative.

A number of open source software groups are working with handset
manufacturers, chip-makers and software developers to coordinate
research into using Linux for mobile devices. They are also trying to
tackle broader issues, says William Weinberg, of Open Source
Development Labs, one of the groups involved. "The handset
manufacturers want their devices to become platforms for innovation. To
do that, they need to expand the number of people who can actually
enhance the devices with software, and for that they need to increase
openness."

tmcnet.com