As computing power increased in the early 1980's, people began to
consider attaching more complex peripheral devices to their computers.
People began to understand the importance of the ``man-machine''
(``person-machine''?) interface. As Rheingold
outlines [Rhe91], early researchers, from Ivan Sutherland
to Richard Bolt at ArchMac
began to realise that the more natural the
interface, the more useful the computer would become.
At the same time, various approaches were being made to improve techniques for computer-based pattern recognition, a task which was easy for humans, yet difficult for computers. There were a number of early failures, due to the huge computing power required to do such tasks. However, a number of approaches developed, such as neural networks (a version of the earlier ``perceptrons''), Bayesian and other statistical methods, instance-based learning algorithms (such as IBL) and symbolic learning packages (such as ID3 and C4.5).
A spin-off of the great advances in interface design was the creation
of ``manual'' input devices. There were of course the very conspicuous
forms, which were to use Sturman's words ``devices for interaction via
the hand and not `whole-hand input' devices''. This
included light-pens, styli, and mice. The first real
``non-conspicuous'' interaction device was the ``Put-that-there''
system, developed at ArchMac by Richard Bolt [Bol80]. This
device used a combination of the newly commercialised Polhemus 6D
tracking device and some simple voice recognition software. On the
screen, the user saw objects. The user would then ``point'' at an
object and say ``Put that ...'' move his/her finger to where he/she
wanted the object to be, and say ``there''.
While this does not connect directly to the matter at
hand, it laid the groundwork for
future developments.