Berkeley – After more than a million years of computation by more
than 4 million computers worldwide, the SETI@home screensaver that
crunches data in search of intelligent signals from space has
produced a list of candidate radio sources that deserve a second look.

Three members of the SETI@home team will head to Puerto Rico this
month to point the Arecibo radio telescope at up to 150 spots
identified as the source of possible signals from intelligent
civilizations.

SETI@home is a computer program disguised as a screen saver that pops
up when a computer is idle and analyzes radio telescope data in
search of strong or unusual signals from space. The candidates for
re-observation are particularly strong signals or ones that have been
observed in the same spot more than once, some of them five or six
times.

“This is the culmination of more than three years of computing, the
largest computation ever done,” said UC Berkeley computer scientist
David Anderson, director of SETI@home. “It’s a milestone for the
SETI@home project.”

SETI@home users should find out the results of the re-observations –
what The Planetary Society, the founding and principal sponsor of
SETI@home, is billing as the “stellar countdown” – within two to
three months.

Though excited at the opportunity to re-observe as many as 150
candidate signals, Anderson is cautious about raising people’s
expectations that they will discover a signal from an
extraterrestrial civilization.

“If there is any possibility at all of finding an extraterrestrial
signal, it’s probably much less than one percent,” he said.

UC Berkeley physicist Dan Werthimer, SETI@home chief scientist, isn’t
getting his hopes up, either. He has conducted a Search for
Extraterrestrial Intelligence (SETI) for 24 years – 11 years using
Arecibo’s 1,000-foot diameter radio dish – and has returned several
times to look again at promising locations and frequency ranges to
determine if a strong radio signal is more than random noise, a
glitch or a passing satellite. He has been disappointed each time.

On the other hand, SETI@home has mobilized so much more computing
power than has ever before been thrown at signal analysis, that the
team has been able to perform much more detailed and complicated
computations on the radio data than now possible with Werthimer’s
ongoing SETI project, called SERENDIP IV (Search for Extraterrestrial
Radio Emissions from Nearby Developed Intelligent Populations).

“I give it a one in 10,000 chance that one of our candidate signals
turns out to be from ET,” said Werthimer, who will head for Puerto
Rico on March 16.

“Whether or not SETI@home succeeds in finding evidence of
extraterrestrial intelligence at this early date, this project has
already made history,” said Bruce Murray, chairman of The Planetary
Society’s board of directors. “SETI@home has performed the most
sensitive and detailed SETI sky survey to date, has demonstrated the
power of the Internet for doing scientific distributed computing, and
has allowed the general public to participate directly in an exciting
research project.”

To acknowledge the 4,287,000-plus users who have analyzed radio data,
the SETI@home team will post on its Web site the names of those
participants who flagged the candidate signals as a result of data
analysis on their home computers. Each candidate signal was analyzed
by several people, because SETI@home sends the same data to more than
one person to double-check results.

The list of candidates is far longer than 150, but Werthimer suspects
that 150 is the maximum he and two colleagues will be able to observe
during the 24 hours total available to them at Arecibo Observatory on
Mar. 18-20. Criteria for inclusion in the list include not only a
strong radio signal and a signal observed more than once in the same
spot and frequency range, but also the signal’s proximity to a known
star and whether that star is known to have planets.

“These factors let us estimate the probability that a candidate is
noise,” Anderson said. “We’re interested in the candidates that are
least likely to be noise.”

Limited analysis of the signals will be done while the team collects
the data, so that observations can be halted and repeated if a very
strong signal reappears. Werthimer will be assisted by graduate
student Paul Demorest and project scientist Eric Korpela.

A more detailed analysis will be conducted later, Anderson said,
ideally with a new version of the SETI@home screensaver based on a
new distributed computing platform called BOINC (Berkeley Open
Infrastructure for Network Computing).

SETI@home offered its screensaver to the world in May 1999 as the
first example of wide-scale distributed computing – linking idle
computers through the Internet to tackle large computational
problems. The key to its success was a fruitful collaboration between
Anderson, a computer scientist who was one of the principal
developers of distributed computing, and Werthimer, a physicist with
two decades of experience collecting radio data and parsing it in
search of unusual signals from space.

Together, they have drawn in not only sci-fi fans and computer geeks,
but many others interested in offering use of their computer to
benefit worthwhile projects. SETI@home has spawned numerous other
distributed computing projects, including Folding@home to calculate
the three-dimensional structure of proteins and climateprediction.net
to improve scientific forecasts of 21st century climate.

However, scientists interested in launching similar projects have
been daunted by the time and money needed to create the software. To
address this problem, Anderson developed BOINC, which is funded by
the National Science Foundation. In addition to being a general
purpose platform, it allows users to partition their computer time
among several distributed computing projects.

“BOINC makes it easy for scientists to set up new SETI@home-type
projects, and to update their applications on the fly,” he said.
“Each change to SETI@home required all our users to download and
install a new program version, but BOINC manages this process without
user intervention.”

BOINC also has the capacity to store data in participants’ unused
disk space, much the way Napster, Gnutella and Kazaa take advantage
of PC hard drives to store MP3 music files.

“The amount of unused disk space out there is staggering,” Anderson
said. “BOINC will let us experiment with new ways of handling data,
like sending it through high-speed Internet connections straight from
telescopes to PCs and archiving it redundantly on PC disks. This will
greatly expand the scope of our SETI research.”

The test case for BOINC is Astropulse, which is designed specifically
to re-examine SETI@home data in search of short radio pulses,
something neither SETI@home nor any other SETI project currently can
do very well. According to Werthimer, Astropulse can detect pulsars,
which blink on and off at periods up to nearly a millisecond;
evaporating black holes, which should emit a brief pulse of radio
waves as they blink out of existence; as well as messages from
extraterrestrials.

“Astropulse will be the first big test of BOINC,” which also provides
enhanced, more realistic 3-D graphics, Anderson said. “If we get
maybe 1,000 people participating in the first BOINC trial, we could
analyze the re-observation data in just a few days.”

“This is a whole new way to look for ET,” Werthimer said.

He and Anderson emphasize that, while the re-observations are the
culmination of nearly four years of data crunching, SETI@home is not
coming to an end. Werthimer hopes to set up a southern hemisphere
SETI program at Parkes Observatory in Australia, the data to be fed
into SETI@home. And data still comes in from the SERENDIP IV
instruments on the Arecibo dish, which will soon use upgraded
receivers to record data from more than one area of the sky at once.

“This is a milestone, but SETI@home will go on,” Anderson said.

In addition to The Planetary Society, other major funders include Sun
Microsystems, the University of California, Quantum Corp., Fujifilm
Computer Products and Network Appliance.

NOTE: David Anderson can be reached at (510) 845-9854 or
davea@ssl.berkeley.edu. Dan Werthimer can be reached at (510)
642-6997, though he will be at Arecibo Observatory from March 16
through March 24. While in Puerto Rico, Werthimer can be reached at
(787) 878-2612 x211, or via e-mail at danw@sunspot.ssl.berkeley.edu.