Information Technology Reference
In-Depth Information
may work fine for a computer conceptually, in practice it has several
problems. First, considerable circuitry is needed to distinguish one volt
age from another. Second, circuits typically yield some voltage loss in
various components, so voltages of 6.6 or 7.3 volts might reasonably oc
cur. In a decimalbased computer, the interpretation of such intermedi
ate voltages adds considerable complexity to circuits. Such complexity
affects costs and reliability, so decimalbased machines can be relatively
difficult and costly to design and maintain.
In a typical binary machine, a common convention is that any
voltage from 0 to 1.3 volts is interpreted as the digit 0, while volt
ages above 1.7 volts represent the digit 1. Having a range of volt
ages allows circuits to accommodate variations of voltage that natu
rally may arise. (Circuits are designed to avoid voltages between 1.3
and 1.7 volts, so such voltages may be considered errors.) With no
other choices regarding possible digit values, processing can utilize
simple and cheap on/off circuits.
Technically, a single 0 or 1 is called a
bit
of information. Such a
data element is sufficient to hold the answer to a yes/no or true/false
question, but a single bit is inadequate to store more general informa
tion. Thus, in a binary system, several bits typically are grouped to
gether in various size units, so that they can hold greater information.
A particularly common grouping involves 8 bits and is called a
byte
.
To see how natural numbers, the integers starting with 0 and
going higher, can be stored in a byte, consider the following num
bering scheme. (If you know about the binary system, this will seem
quite familiar.) Note that there are eight columns, or place values,
because this is an 8bit number. The placement of each 0 and 1 in
fluences their value. To convert a binary number into a decimal
number, start at the right of a sequence of 0's and 1's, and label the
digits with a power of 2. Then add the labels corresponding to
where the digit 1 appears in the sequence. Here is an example:
8bit number:
0 1 0 0 1 1 0 1
↑ ↑↑↑↑↑↑↑
Label:
128
64
32
16
8421
(or power of 2):
2
7
2
6
2
5
2
4
2
3
2
2
2
1
2
0
Number's value: 77
64
8
4
1
