38 | New Scientist | 25 July 2020
to chart their change. By operating on the
numbers independently of how the physical
equipment itself works, digital computing
can be hugely versatile. Swapping between
different problems just means setting up a
different set of mathematical commands,
rather than re-engineering cogs or their
equivalent for each type of calculation.A little bit better
A key characteristic of digital computers is
the use of binary digits – also known as bits –
that represent all the processed or stored
data as a string of 0s and 1s. In the first digital
computers, information was stored and fed
in via punch cards with holes representing
0s and solid card representing 1s. For the
calculation itself, the computers read the
information and translate it onto circuits
equipped with transistors capable of
switching between two states – that is,
routing a current one way or another.
Processing the data then involves following
a program that flips the right set of switches
at each stage of the calculation.
One drawback of storing data in the form
of binary digits is that the values for variables
are no longer continuous. Whereas a pointer
on a cog can rotate seamlessly through
everything between the numbers 4 and 5,
for instance, a basic digital computer might
jump from 4.1 to 4.2 without being able to
represent the values in between. Adding
more bits can make the gaps between
numbers ever smaller, but having to
make jumps of some size is inevitable.
This doesn’t necessarily mean a drop
in accuracy. Think of a digital stopwatch
versus an analogue one: although a precise
measurement of the angle of the second
hand could, in theory, allow you to read off
infinitely small units of time, in reality the
eye won’t read a simple dial with the same
accuracy that digital figures can drill down to.
The first programmable, general-purpose
digital computer was the Electronic
Numerical Integrator and Computer (ENIAC),
introduced in 1946. It was the size of a roomLOUISA^ GOULIAMAKI/AFP
VIAGETTY^ IMAGESand took days to program, but it was
significantly more powerful than any
computer that had come before. Analogue
approaches kept up for a while, but they
were little more than a quaint memory
by the 1980s.
Yet even today the world isn’t as digital
as it seems. “The physical world is analogue,”
says Yannis Tsividis at Columbia University
in New York. Analogue technology is still all
around us. The electromagnetic radio signals
that our smartphones use to communicate
with each other, for instance, are analogue,
requiring analogue-to-digital converters
to allow the phone’s digital electronics to
process them.
Analogue isn’t just useful for shifting
data from one place to another. There
are situations where these technologies
are proving superior to digital ones for
processing data, too. One key area concerns
the kinds of equations used for everything
from modelling the effects of hormone levels
in the body to understanding particle
behaviour. These differential and integralThe Antikythera mechanism, an
analogue computer from ancient
Greece (above); IBM’s Blue Gene
digital supercomputer (below)EVER
ET
T^ COL
LE
CT
IONHISTORICAL
/ALAMY