11
A comparison of Figures 5 and 8 might suggest that we have progressed
in a circle. In both cases the servo loop is closed inside a computer.
However, in reality the common trends are again verified: the trend has
been from specific to general, and the presence of a microprocessor in
the servo- system is a good example of distributed processing.
Analytical Plotters
When discussing here the design trends in analytical plotters there
is no reason to start with a historical perspective. Much of what was
said previously about recording systems and servos applies to input and
output functions of analytical plotters as their development has progressed
over the years. Let us therefore start from the present. Figure 9 shows
a typical design approach that may be applied to analytical plotters today.
All minicomputers in the class capable of handling the analytical plotter
requirements have an I/O bus. Peripheral equipments communicate with the
computer via this bus, to which they are connected by a bus interface.
The bus ipterface contains all the necessary timing and control functions
to establish the communication link, as well as pertinent address and
data transfer contacts. The input and output functions of an analytical
plotter can be tied to the computer by interfacing to its I/O bus. Fig
ure 9 shows schematically what is involved. Essentially, there is a
trade-off between the number of bus interface units employed and the
amount of additional control logic. A practical and cost-effective com
promise in most cases is to use several bus interface units, as shown in
the figure.
Microprocessors are inherently capable of performing various logic
functions required in interfacing with the computer I/O bus, and with
other devices. Specifically, they can multiplex data in and out, and
between these functions also perform data accumulation or "metering out"
for servos. Figure 10 shows a possible design approach suitable for
the relatively slow microprocessors readily available today.
The ever-increasing speed of microprocessors leads to another design
concept that may in the near future become feasible. As shown in Fig
ure 11, all input/output functions are channelled through a single micro
processor. Obviously, in line of what was said about distributed pro
cessing and servos, the input and servo channels may include dedicated
microprocessors of their own, as means of implementing necessary logic
functions.
Finally, the computer may become a microcomputer, as shown in Fig
ure 12. It may seem again that we have progressed in a circle. This
is not so: the system of Figure 12 is a distributed processing system
with a hierarchy of processors. Such a concept is certainly intellec
tually pleasing; its significance, however, is in equal measure based
on practicality. All the microprocessors may be identical, differing