25 July 2020 | New Scientist | 37A
MID an epic haul of jewels and
statues rescued from an ancient
Greek shipwreck, it was a lump
of corroded wood and bronze that would
capture the world’s imagination. Pulled
from the Mediterranean Sea in 1901, the
Antikythera mechanism was an astonishingly
sophisticated 2000-year-old computer.
The size of a shoebox, sporting chunky
bronze gears rimmed with hundreds of
triangular teeth, it was built to chart the
paths of celestial bodies and was capable
of addition, multiplication, subtraction
and division – all by cranking a handle.
Nearly half a century would pass before
its significance became apparent. By that
time, the world had developed a new breed
of calculators: the digital computers we still
use today. Powered by electricity rather than
a hand crank, they were a huge step forward
and have proved fast and powerful enough
to keep the modern world afloat.
But digital computers aren’t always
the best tool for the job. Much of the
mathematics used at the frontiers of
modern science, for example, translates
awkwardly into digital technology, where
TOcertain equations are cumbersome to solve. >
TT
O^ R
EN
NA
Analogue
comeback
Old-school technology may shape the future
of digital computing, finds Anna Demming
New approaches are increasingly sought
after in the design of artificial intelligence,
too, where digital computers struggle to
mimic the complex processes of the human
brain. As such, the latest hardware is often too
expensive and inefficient to use in this area.
Mechanical devices are unlikely to be
the answer. Even so, to better meet today’s
challenges, researchers at the cutting edge
of computer development are looking to
analogue techniques that have more in
common with the Antikythera mechanism
than today’s conventional computers. To
save computing’s future, we may need a
blast from the past.
As the name suggests, analogue computers
can provide a physical analogue of the system
they are describing. In a device like the
Antikythera mechanism that is made up of a
set of rotating cogs, for example, the positions
of certain cogs represent the locations of the
sun and moon. You can always tell where
those objects are just by looking at the cogs.
By contrast, a digital machine has no
such exact mappings. Instead, it converts
all information to numbers, such as the
coordinates of the sun and moon, and
performs calculations on those numbers