Representation of Integers and Floating Points in Computers
Representation of Integers and Floating Points in Computers
A Tutorial on Data
1.1 Decimal (Base 10) Number Sys
1.2 Binary (Base 2) Number System
1.3 Hexadecimal (Base 16) Numbe
Representation 1.4
1.5
Conversion from Hexadecimal
Conversion from Binary to Hex
1.6 Conversion from Base r to De
Integers, Floating-point Numbers, 1.7 Conversion from Decimal (Bas
1.8 Conversion between Two Num
and Characters 1.9 Exercises (Number Systems Co
2. Computer Memory & Data Repr
3. Integer Representation
3.1 n-bit Unsigned Integers
3.2 Signed Integers
1. Number Systems 3.3 n-bit Sign Integers in Sign-Ma
3.4 n-bit Sign Integers in 1's Comp
Human beings use decimal (base 10) and duodecimal (base 12) number systems for counting and 3.5 n-bit Sign Integers in 2's Comp
measurements (probably because we have 10 fingers and two big toes). Computers use binary (base 2) 3.6 Computers use 2's Compleme
number system, as they are made from binary digital components (known as transistors) operating in 3.7 Range of n-bit 2's Complemen
two states - on and off. In computing, we also use hexadecimal (base 16) or octal (base 8) number 3.8 Decoding 2's Complement Nu
systems, as a compact form for representing binary numbers. 3.9 Big Endian vs. Little Endian
3.10 Exercise (Integer Representat
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 1/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
735 = 700 + 30 + 5 = 7×10^2 + 3×10^1 + 5×10^0 5.1 7-bit ASCII Code (aka US-ASC
5.2 8-bit Latin-1 (aka ISO/IEC 8859
We shall denote a decimal number with an optional suffix D if ambiguity arises. 5.3 Other 8-bit Extension of US-A
5.4 Unicode (aka ISO/IEC 10646 U
5.5 UTF-8 (Unicode Transformatio
1.2 Binary (Base 2) Number System
5.6 UTF-16 (Unicode Transformati
Binary number system has two symbols: 0 and 1, called bits. It is also a positional notation, for example, 5.7 UTF-32 (Unicode Transformati
5.8 Formats of Multi-Byte (e.g., Un
10110B = 10000B + 0000B + 100B + 10B + 0B = 1×2^4 + 0×2^3 + 1×2^2 + 1×2^1 + 0×2^0
5.9 Formats of Text Files
We shall denote a binary number with a suffix B. Some programming languages denote binary numbers 5.10 Windows' CMD Codepage
with prefix 0b or 0B (e.g., 0b1001000), or prefix b with the bits quoted (e.g., b'10001111'). 5.11 Chinese Character Sets
5.12 Collating Sequences (for Ran
A binary digit is called a bit. Eight bits is called a byte (why 8-bit unit? Probably because 8=23). 5.13 For Java Programmers - java
5.14 For Java Programmers - char
1.3 Hexadecimal (Base 16) Number System 5.15 Displaying Hex Values & Hex
6. Summary - Why Bother about D
Hexadecimal number system uses 16 symbols: 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, A, B, C, D, E, and F, called hex
6.1 Exercises (Data Representation
digits. It is a positional notation, for example,
We shall denote a hexadecimal number (in short, hex) with a suffix H. Some programming languages denote hex numbers with prefix 0x or 0X (e.g.,
0x1A3C5F), or prefix x with hex digits quoted (e.g., x'C3A4D98B').
Each hexadecimal digit is also called a hex digit. Most programming languages accept lowercase 'a' to 'f' as well as uppercase 'A' to 'F'.
Computers uses binary system in their internal operations, as they are built from binary digital electronic components with 2 states - on and off.
However, writing or reading a long sequence of binary bits is cumbersome and error-prone (try to read this binary string: 1011 0011 0100 0011
0001 1101 0001 1000B, which is the same as hexadecimal B343 1D18H). Hexadecimal system is used as a compact form or shorthand for binary
bits. Each hex digit is equivalent to 4 binary bits, i.e., shorthand for 4 bits, as follows:
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 2/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
4 0100 4
5 0101 5
6 0110 6
7 0111 7
8 1000 8
9 1001 9
A 1010 10
B 1011 11
C 1100 12
D 1101 13
E 1110 14
F 1111 15
It is important to note that hexadecimal number provides a compact form or shorthand for representing binary bits.
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 3/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
For examples,
The above procedure is actually applicable to conversion between any 2 base systems. For example,
c. 101010101010
3. Convert the following hexadecimal numbers into binary and decimal numbers:
a. ABCDE
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 5/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
b. 1234
c. 80F
4. Convert the following decimal numbers into binary equivalent:
a. 19.25D
b. 123.456D
Answers: You could use the Windows' Calculator (calc.exe) to carry out number system conversion, by setting it to the Programmer or scientific
mode. (Run "calc" ⇒ Select "Settings" menu ⇒ Choose "Programmer" or "Scientific" mode.)
1. 1101100B, 1001011110000B, 10001100101000B, 6CH, 12F0H, 2328H.
Integers, for example, can be represented in 8-bit, 16-bit, 32-bit or 64-bit. You, as the programmer, choose an appropriate bit-length for your
integers. Your choice will impose constraint on the range of integers that can be represented. Besides the bit-length, an integer can be represented
in various representation schemes, e.g., unsigned vs. signed integers. An 8-bit unsigned integer has a range of 0 to 255, while an 8-bit signed
integer has a range of -128 to 127 - both representing 256 distinct numbers.
It is important to note that a computer memory location merely stores a binary pattern. It is entirely up to you, as the programmer, to decide on
how these patterns are to be interpreted. For example, the 8-bit binary pattern "0100 0001B" can be interpreted as an unsigned integer 65, or an
ASCII character 'A', or some secret information known only to you. In other words, you have to first decide how to represent a piece of data in a
binary pattern before the binary patterns make sense. The interpretation of binary pattern is called data representation or encoding. Furthermore, it
is important that the data representation schemes are agreed-upon by all the parties, i.e., industrial standards need to be formulated and straightly
followed.
Once you decided on the data representation scheme, certain constraints, in particular, the precision and range will be imposed. Hence, it is
important to understand data representation to write correct and high-performance programs.
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 6/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
The moral of the story is unless you know the encoding scheme, there is
no way that you can decode the data.
3. Integer Representation
Integers are whole numbers or fixed-point numbers with the radix point fixed after the least-significant bit. They are contrast to real numbers or
floating-point numbers, where the position of the radix point varies. It is important to take note that integers and floating-point numbers are
treated differently in computers. They have different representation and are processed differently (e.g., floating-point numbers are processed in a
so-called floating-point processor). Floating-point numbers will be discussed later.
Computers use a fixed number of bits to represent an integer. The commonly-used bit-lengths for integers are 8-bit, 16-bit, 32-bit or 64-bit.
Besides bit-lengths, there are two representation schemes for integers:
1. Unsigned Integers: can represent zero and positive integers.
2. Signed Integers: can represent zero, positive and negative integers. Three representation schemes had been proposed for signed integers:
a. Sign-Magnitude representation
b. 1's Complement representation
c. 2's Complement representation
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 7/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
You, as the programmer, need to decide on the bit-length and representation scheme for your integers, depending on your application's
requirements. Suppose that you need a counter for counting a small quantity from 0 up to 200, you might choose the 8-bit unsigned integer
scheme as there is no negative numbers involved.
Example 1: Suppose that n=8 and the binary pattern is 0100 0001B, the value of this unsigned integer is 1×2^0 + 1×2^6 = 65D.
Example 2: Suppose that n=16 and the binary pattern is 0001 0000 0000 1000B, the value of this unsigned integer is 1×2^3 + 1×2^12 =
4104D.
Example 3: Suppose that n=16 and the binary pattern is 0000 0000 0000 0000B, the value of this unsigned integer is 0.
An n-bit pattern can represent 2^n distinct integers. An n-bit unsigned integer can represent integers from 0 to (2^n)-1, as tabulated below:
n Minimum Maximum
8 0 (2^8)-1 (=255)
16 0 (2^16)-1 (=65,535)
32 0 (2^32)-1 (=4,294,967,295) (9+ digits)
64 0 (2^64)-1 (=18,446,744,073,709,551,615) (19+ digits)
In all the above three schemes, the most-significant bit (msb) is called the sign bit. The sign bit is used to represent the sign of the integer - with 0
for positive integers and 1 for negative integers. The magnitude of the integer, however, is interpreted differently in different schemes.
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 8/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
In sign-magnitude representation:
The most-significant bit (msb) is the sign bit, with value of 0 representing positive integer and 1 representing negative integer.
The remaining n-1 bits represents the magnitude (absolute value) of the integer. The absolute value of the integer is interpreted as "the
magnitude of the (n-1)-bit binary pattern".
Example 1 : Suppose that n=8 and the binary representation is 0 100 0001B.
Sign bit is 0 ⇒ positive
Absolute value is 100 0001B = 65D
Hence, the integer is +65D
Example 2 : Suppose that n=8 and the binary representation is 1 000 0001B.
Sign bit is 1 ⇒ negative
Absolute value is 000 0001B = 1D
Hence, the integer is -1D
Example 3 : Suppose that n=8 and the binary representation is 0 000 0000B.
Sign bit is 0 ⇒ positive
Absolute value is 000 0000B = 0D
Hence, the integer is +0D
Example 4 : Suppose that n=8 and the binary representation is 1 000 0000B.
Sign bit is 1 ⇒ negative
Absolute value is 000 0000B = 0D
Hence, the integer is -0D
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 9/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
Example 1 : Suppose that n=8 and the binary representation 0 100 0001B.
Sign bit is 0 ⇒ positive
Absolute value is 100 0001B = 65D
Hence, the integer is +65D
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 10/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
Example 2 : Suppose that n=8 and the binary representation 1 000 0001B.
Sign bit is 1 ⇒ negative
Absolute value is the complement of 000 0001B, i.e., 111 1110B = 126D
Hence, the integer is -126D
Example 3 : Suppose that n=8 and the binary representation 0 000 0000B.
Sign bit is 0 ⇒ positive
Absolute value is 000 0000B = 0D
Hence, the integer is +0D
Example 4 : Suppose that n=8 and the binary representation 1 111 1111B.
Sign bit is 1 ⇒ negative
Absolute value is the complement of 111 1111B, i.e., 000 0000B = 0D
Hence, the integer is -0D
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 11/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
Example 1 : Suppose that n=8 and the binary representation 0 100 0001B.
Sign bit is 0 ⇒ positive
Absolute value is 100 0001B = 65D
Hence, the integer is +65D
Example 2 : Suppose that n=8 and the binary representation 1 000 0001B.
Sign bit is 1 ⇒ negative
Absolute value is the complement of 000 0001B plus 1, i.e., 111 1110B + 1B = 127D
Hence, the integer is -127D
Example 3 : Suppose that n=8 and the binary representation 0 000 0000B.
Sign bit is 0 ⇒ positive
Absolute value is 000 0000B = 0D
Hence, the integer is +0D
Example 4 : Suppose that n=8 and the binary representation 1 111 1111B.
Sign bit is 1 ⇒ negative
Absolute value is the complement of 111 1111B plus 1, i.e., 000 0000B + 1B = 1D
Hence, the integer is -1D
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 12/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 13/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
We have discussed three representations for signed integers: signed-magnitude, 1's complement and 2's complement. Computers use 2's
complement in representing signed integers. This is because:
1. There is only one representation for the number zero in 2's complement, instead of two representations in sign-magnitude and 1's
complement.
2. Positive and negative integers can be treated together in addition and subtraction. Subtraction can be carried out using the "addition logic".
Example 1: Addition of Two Positive Integers: Suppose that n=8, 65D + 5D = 70D
Example 2: Subtraction is treated as Addition of a Positive and a Negative Integers: Suppose that n=8, 5D - 5D = 65D
+ (-5D) = 60D
Example 3: Addition of Two Negative Integers: Suppose that n=8, -65D - 5D = (-65D) + (-5D) = -70D
Because of the fixed precision (i.e., fixed number of bits), an n-bit 2's complement signed integer has a certain range. For example, for n=8, the range
of 2's complement signed integers is -128 to +127. During addition (and subtraction), it is important to check whether the result exceeds this
range, in other words, whether overflow or underflow has occurred.
Example 4: Overflow: Suppose that n=8, 127D + 2D = 129D (overflow - beyond the range)
Example 5: Underflow: Suppose that n=8, -125D - 5D = -130D (underflow - below the range)
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 14/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
The following diagram explains how the 2's complement works. By re-arranging the number line, values from -128 to +127 are represented
contiguously by ignoring the carry bit.
n minimum maximum
8 -(2^7) (=-128) +(2^7)-1 (=+127)
16 -(2^15) (=-32,768) +(2^15)-1 (=+32,767)
32 -(2^31) (=-2,147,483,648) +(2^31)-1 (=+2,147,483,647)(9+ digits)
64 -(2^63) +(2^63)-1 (=+9,223,372,036,854,775,807)(18+
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 15/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
(=-9,223,372,036,854,775,808) digits)
3. If S=1, the number is negative. you could "invert the n-1 bits and plus 1" to get the absolute value of negative number.
Alternatively, you could scan the remaining n-1 bits from the right (least-significant bit). Look for the first occurrence of 1. Flip all the bits to
the left of that first occurrence of 1. The flipped pattern gives the absolute value. For example,
The term"Endian" refers to the order of storing bytes in computer memory. In "Big Endian" scheme, the most significant byte is stored first in the
lowest memory address (or big in first), while "Little Endian" stores the least significant bytes in the lowest memory address.
For example, the 32-bit integer 12345678H (30541989610) is stored as 12H 34H 56H 78H in big endian; and 78H 56H 34H 12H in little endian. An
16-bit integer 00H 01H is interpreted as 0001H in big endian, and 0100H as little endian.
4. Give the value of +88, -88 , -1, 0, +1, -127, and +127 in 8-bit sign-magnitude representation.
5. Give the value of +88, -88 , -1, 0, +1, -127 and +127 in 8-bit 1's complement representation.
6. [TODO] more.
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 16/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
Answers
1. The range of unsigned n-bit integers is [0, 2^n - 1]. The range of n-bit 2's complement signed integer is [-2^(n-1), +2^(n-1)-1];
2. 88 (0101 1000), 0 (0000 0000), 1 (0000 0001), 127 (0111 1111), 255 (1111 1111).
3. +88 (0101 1000), -88 (1010 1000), -1 (1111 1111), 0 (0000 0000), +1 (0000 0001), -128 (1000 0000), +127 (0111 1111).
4. +88 (0101 1000), -88 (1101 1000), -1 (1000 0001), 0 (0000 0000 or 1000 0000), +1 (0000 0001), -127 (1111 1111), +127
(0111 1111).
5. +88 (0101 1000), -88 (1010 0111), -1 (1111 1110), 0 (0000 0000 or 1111 1111), +1 (0000 0001), -127 (1000 0000), +127
(0111 1111).
A floating-point number is typically expressed in the scientific notation, with a fraction (F), and an exponent (E) of a certain radix (r), in the form of
F×r^E. Decimal numbers use radix of 10 (F×10^E); while binary numbers use radix of 2 (F×2^E).
Representation of floating point number is not unique. For example, the number 55.66 can be represented as 5.566×10^1, 0.5566×10^2,
0.05566×10^3, and so on. The fractional part can be normalized. In the normalized form, there is only a single non-zero digit before the radix
point. For example, decimal number 123.4567 can be normalized as 1.234567×10^2; binary number 1010.1011B can be normalized as
1.0101011B×2^3.
It is important to note that floating-point numbers suffer from loss of precision when represented with a fixed number of bits (e.g., 32-bit or 64-bit).
This is because there are infinite number of real numbers (even within a small range of says 0.0 to 0.1). On the other hand, a n-bit binary pattern
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 17/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
can represent a finite 2^n distinct numbers. Hence, not all the real numbers can be represented. The nearest approximation will be used instead,
resulted in loss of accuracy.
It is also important to note that floating number arithmetic is very much less efficient than integer arithmetic. It could be speed up with a so-called
dedicated floating-point co-processor. Hence, use integers if your application does not require floating-point numbers.
In computers, floating-point numbers are represented in scientific notation of fraction (F) and exponent (E) with a radix of 2, in the form of F×2^E.
Both E and F can be positive as well as negative. Modern computers adopt IEEE 754 standard for representing floating-point numbers. There are
two representation schemes: 32-bit single-precision and 64-bit double-precision.
Normalized Form
Let's illustrate with an example, suppose that the 32-bit pattern is 1 1000 0001 011 0000 0000 0000 0000 0000, with:
S = 1
E = 1000 0001
F = 011 0000 0000 0000 0000 0000
In the normalized form, the actual fraction is normalized with an implicit leading 1 in the form of 1.F. In this example, the actual fraction is 1.011
0000 0000 0000 0000 0000 = 1 + 1×2^-2 + 1×2^-3 = 1.375D.
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 18/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
The sign bit represents the sign of the number, with S=0 for positive and S=1 for negative number. In this example with S=1, this is a negative
number, i.e., -1.375D.
In normalized form, the actual exponent is E-127 (so-called excess-127 or bias-127). This is because we need to represent both positive and
negative exponent. With an 8-bit E, ranging from 0 to 255, the excess-127 scheme could provide actual exponent of -127 to 128. In this example,
E-127=129-127=2D.
De-Normalized Form
Normalized form has a serious problem, with an implicit leading 1 for the fraction, it cannot represent the number zero! Convince yourself on this!
For E=0, the numbers are in the de-normalized form. An implicit leading 0 (instead of 1) is used for the fraction; and the actual exponent is always
-126. Hence, the number zero can be represented with E=0 and F=0 (because 0.0×2^-126=0).
We can also represent very small positive and negative numbers in de-normalized form with E=0. For example, if S=1, E=0, and F=011 0000 0000
0000 0000 0000. The actual fraction is 0.011=1×2^-2+1×2^-3=0.375D. Since S=1, it is a negative number. With E=0, the actual exponent is -126.
Hence the number is -0.375×2^-126 = -4.4×10^-39, which is an extremely small negative number (close to zero).
Summar y
In summary, the value (N) is calculated as follows:
For 1 ≤ E ≤ 254, N = (-1)^S × 1.F × 2^(E-127). These numbers are in the so-called normalized form. The sign-bit represents the sign
of the number. Fractional part (1.F) are normalized with an implicit leading 1. The exponent is bias (or in excess) of 127, so as to represent
both positive and negative exponent. The range of exponent is -126 to +127.
For E = 0, N = (-1)^S × 0.F × 2^(-126). These numbers are in the so-called denormalized form. The exponent of 2^-126 evaluates to a
very small number. Denormalized form is needed to represent zero (with F=0 and E=0). It can also represents very small positive and negative
number close to zero.
For E = 255, it represents special values, such as ±INF (positive and negative infinity) and NaN (not a number). This is beyond the scope of this
article.
Example 1: Suppose that IEEE-754 32-bit floating-point representation pattern is 0 10000000 110 0000 0000 0000 0000 0000.
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 19/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
Example 2: Suppose that IEEE-754 32-bit floating-point representation pattern is 1 01111110 100 0000 0000 0000 0000 0000.
Example 3: Suppose that IEEE-754 32-bit floating-point representation pattern is 1 01111110 000 0000 0000 0000 0000 0001.
Example 4 (De-Normalized Form): Suppose that IEEE-754 32-bit floating-point representation pattern is 1 00000000 000 0000 0000
0000 0000 0001.
Hints:
1. Largest positive number: S=0, E=1111 1110 (254), F=111 1111 1111 1111 1111 1111.
Smallest positive number: S=0, E=0000 00001 (1), F=000 0000 0000 0000 0000 0000.
2. Same as above, but S=1.
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 20/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
3. Largest positive number: S=0, E=0, F=111 1111 1111 1111 1111 1111.
Smallest positive number: S=0, E=0, F=000 0000 0000 0000 0000 0001.
4. Same as above, but S=1.
System.out.println(Float.intBitsToFloat(0x7fffff));
System.out.println(Double.longBitsToDouble(0x1fffffffffffffL));
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 21/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
Take note that n-bit pattern has a finite number of combinations (=2^n), which could represent finite distinct numbers. It is not possible to
represent the infinite numbers in the real axis (even a small range says 0.0 to 1.0 has infinite numbers). That is, not all floating-point numbers can
be accurately represented. Instead, the closest approximation is used, which leads to loss of accuracy.
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 22/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
For single-precision, E = 0,
N = (-1)^S × 0.F × 2^(-126)
For double-precision, E = 0,
N = (-1)^S × 0.F × 2^(-1022)
Denormalized form can represent very small numbers closed to zero, and zero, which cannot be represented in normalized form, as shown in the
above figure.
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 23/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
Special Values
Zero: Zero cannot be represented in the normalized form, and must be represented in denormalized form with E=0 and F=0. There are two
representations for zero: +0 with S=0 and -0 with S=1.
Infinity: The value of +infinity (e.g., 1/0) and -infinity (e.g., -1/0) are represented with an exponent of all 1's (E = 255 for single-precision and E =
2047 for double-precision), F=0, and S=0 (for +INF) and S=1 (for -INF).
Not a Number (NaN): NaN denotes a value that cannot be represented as real number (e.g. 0/0). NaN is represented with Exponent of all 1's (E =
255 for single-precision and E = 2047 for double-precision) and any non-zero fraction.
5. Character Encoding
In computer memory, character are "encoded" (or "represented") using a chosen "character encoding schemes" (aka "character set", "charset",
"character map", or "code page").
For example, in ASCII (as well as Latin1, Unicode, and many other character sets):
code numbers 65D (41H) to 90D (5AH) represents 'A' to 'Z', respectively.
code numbers 97D (61H) to 122D (7AH) represents 'a' to 'z', respectively.
code numbers 48D (30H) to 57D (39H) represents '0' to '9', respectively.
It is important to note that the representation scheme must be known before a binary pattern can be interpreted. E.g., the 8-bit pattern "0100
0010B" could represent anything under the sun known only to the person encoded it.
The most commonly-used character encoding schemes are: 7-bit ASCII (ISO/IEC 646) and 8-bit Latin-x (ISO/IEC 8859-x) for western european
characters, and Unicode (ISO/IEC 10646) for internationalization (i18n).
A 7-bit encoding scheme (such as ASCII) can represent 128 characters and symbols. An 8-bit character encoding scheme (such as Latin-x) can
represent 256 characters and symbols; whereas a 16-bit encoding scheme (such as Unicode UCS-2) can represents 65,536 characters and symbols.
5.1 7-bit ASCII Code (aka US-ASCII, ISO/IEC 646, ITU-T T.50)
ASCII (American Standard Code for Information Interchange) is one of the earlier character coding schemes.
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 24/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
ASCII is originally a 7-bit code. It has been extended to 8-bit to better utilize the 8-bit computer memory organization. (The 8th-bit was
originally used for parity check in the early computers.)
Code numbers 32D (20H) to 126D (7EH) are printable (displayable) characters as tabulated (arranged in hexadecimal and decimal) as follows:
Hex 0 1 2 3 4 5 6 7 8 9 A B C D E F
3 0 1 2 3 4 5 6 7 8 9 : ; < = > ?
4 @ A B C D E F G H I J K L M N O
5 P Q R S T U V W X Y Z [ \ ] ^ _
6 ` a b c d e f g h i j k l m n o
7 p q r s t u v w x y z { | } ~
Dec 0 1 2 3 4 5 6 7 8 9
4 ( ) * + , - . / 0 1
5 2 3 4 5 6 7 8 9 : ;
6 < = > ? @ A B C D E
7 F G H I J K L M N O
8 P Q R S T U V W X Y
9 Z [ \ ] ^ _ ` a b c
10 d e f g h i j k l m
11 n o p q r s t u v w
12 x y z { | } ~
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 25/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
'A' to 'Z': 41H-5AH (0101 0001B to 0101 1010B) or (010x xxxxB). 'A' to 'Z' are continuous without gap.
'a' to 'z': 61H-7AH (0110 0001B to 0111 1010B) or (011x xxxxB). 'A' to 'Z' are also continuous without gap. However, there is a
gap between uppercase and lowercase letters. To convert between upper and lowercase, flip the value of bit-5.
Code numbers 0D (00H) to 31D (1FH), and 127D (7FH) are special control characters, which are non-printable (non-displayable), as
tabulated below. Many of these characters were used in the early days for transmission control (e.g., STX, ETX) and printer control (e.g., Form-
Feed), which are now obsolete. The remaining meaningful codes today are:
09H for Tab ('\t').
0AH for Line-Feed or newline (LF or '\n') and 0DH for Carriage-Return (CR or 'r'), which are used as line delimiter (aka line separator, end-
of-line) for text files. There is unfortunately no standard for line delimiter: Unixes and Mac use 0AH (LF or "\n"), Windows use 0D0AH (CR+LF
or "\r\n"). Programming languages such as C/C++/Java (which was created on Unix) use 0AH (LF or "\n").
In programming languages such as C/C++/Java, line-feed (0AH) is denoted as '\n', carriage-return (0DH) as '\r', tab (09H) as '\t'.
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 26/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
'\t'
10 0A LF Line Feed '\n' 27 1B ESC Escape
11 0B VT Vertical Feed 28 1C IS4 File
Separator
12 0C FF Form Feed 'f' 29 1D IS3 Group
Separator
13 0D CR Carriage Return 30 1E IS2 Record
'\r' Separator
14 0E SO Shift Out 31 1F IS1 Unit
Separator
15 0F SI Shift In
16 10 DLE Datalink Escape 127 7F DEL Delete
ISO/IEC 8859-1, aka Latin alphabet No. 1, or Latin-1 in short, is the most commonly-used encoding scheme for western european languages. It has
191 printable characters from the latin script, which covers languages like English, German, Italian, Portuguese and Spanish. Latin-1 is backward
compatible with the 7-bit US-ASCII code. That is, the first 128 characters in Latin-1 (code numbers 0 to 127 (7FH)), is the same as US-ASCII. Code
numbers 128 (80H) to 159 (9FH) are not assigned. Code numbers 160 (A0H) to 255 (FFH) are given as follows:
Hex 0 1 2 3 4 5 6 7 8 9 A B C D E F
A NBSP ¡ ¢ £ ¤ ¥ ¦ § ¨ © ª « ¬ SHY ® ¯
B ° ± ² ³ ´ µ ¶ · ¸ ¹ º » ¼ ½ ¾ ¿
C À Á Â Ã Ä Å Æ Ç È É Ê Ë Ì Í Î Ï
D Ð Ñ Ò Ó Ô Õ Ö × Ø Ù Ú Û Ü Ý Þ ß
E à á â ã ä å æ ç è é ê ë ì í î ï
F ð ñ ò ó ô õ ö ÷ ø ù ú û ü ý þ ÿ
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 27/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
ISO/IEC-8859 has 16 parts. Besides the most commonly-used Part 1, Part 2 is meant for Central European (Polish, Czech, Hungarian, etc), Part 3 for
South European (Turkish, etc), Part 4 for North European (Estonian, Latvian, etc), Part 5 for Cyrillic, Part 6 for Arabic, Part 7 for Greek, Part 8 for
Hebrew, Part 9 for Turkish, Part 10 for Nordic, Part 11 for Thai, Part 12 was abandon, Part 13 for Baltic Rim, Part 14 for Celtic, Part 15 for French,
Finnish, etc. Part 16 for South-Eastern European.
ANSI (American National Standards Institute) (aka Windows-1252, or Windows Codepage 1252): for Latin alphabets used in the legacy
DOS/Windows systems. It is a superset of ISO-8859-1 with code numbers 128 (80H) to 159 (9FH) assigned to displayable characters, such as
"smart" single-quotes and double-quotes. A common problem in web browsers is that all the quotes and apostrophes (produced by "smart
quotes" in some Microsoft software) were replaced with question marks or some strange symbols. It it because the document is labeled as ISO-
8859-1 (instead of Windows-1252), where these code numbers are undefined. Most modern browsers and e-mail clients treat charset ISO-8859-1
as Windows-1252 in order to accommodate such mis-labeling.
Hex 0 1 2 3 4 5 6 7 8 9 A B C D E F
8 € ‚ ƒ „ … † ‡ ˆ ‰ Š ‹ Œ Ž
9 ‘ ’ “ ” • – — ™ š › œ ž Ÿ
EBCDIC (Extended Binary Coded Decimal Interchange Code): Used in the early IBM computers.
Unicode aims to provide a standard character encoding scheme, which is universal, efficient, uniform and unambiguous. Unicode standard is
maintained by a non-profit organization called the Unicode Consortium (@ www.unicode.org). Unicode is an ISO/IEC standard 10646.
Unicode is backward compatible with the 7-bit US-ASCII and 8-bit Latin-1 (ISO-8859-1). That is, the first 128 characters are the same as US-ASCII;
and the first 256 characters are the same as Latin-1.
Unicode originally uses 16 bits (called UCS-2 or Unicode Character Set - 2 byte), which can represent up to 65,536 characters. It has since been
expanded to more than 16 bits, currently stands at 21 bits. The range of the legal codes in ISO/IEC 10646 is now from U+0000H to U+10FFFFH (21
bits or about 2 million characters), covering all current and ancient historical scripts. The original 16-bit range of U+0000H to U+FFFFH (65536
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 28/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
characters) is known as Basic Multilingual Plane (BMP), covering all the major languages in use currently. The characters outside BMP are called
Supplementary Characters, which are not frequently-used.
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 29/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
In UTF-8, Unicode numbers corresponding to the 7-bit ASCII characters are padded with a leading zero; thus has the same value as ASCII. Hence,
UTF-8 can be used with all software using ASCII. Unicode numbers of 128 and above, which are less frequently used, are encoded using more bytes
(2-4 bytes). UTF-8 generally requires less storage and is compatible with ASCII. The drawback of UTF-8 is more processing power needed to unpack
the code due to its variable length. UTF-8 is the most popular format for Unicode.
Notes:
UTF-8 uses 1-3 bytes for the characters in BMP (16-bit), and 4 bytes for supplementary characters outside BMP (21-bit).
The 128 ASCII characters (basic Latin letters, digits, and punctuation signs) use one byte. Most European and Middle East characters use a 2-
byte sequence, which includes extended Latin letters (with tilde, macron, acute, grave and other accents), Greek, Armenian, Hebrew, Arabic, and
others. Chinese, Japanese and Korean (CJK) use three-byte sequences.
All the bytes, except the 128 ASCII characters, have a leading '1' bit. In other words, the ASCII bytes, with a leading '0' bit, can be identified
and decoded easily.
Take note that for the 65536 characters in BMP, the UTF-16 is the same as UCS-2 (2 bytes). However, 4 bytes are used for the supplementary
characters outside the BMP.
For BMP characters, UTF-16 is the same as UCS-2. For supplementary characters, each character requires a pair 16-bit values, the first from the
high-surrogates range, (\uD800-\uDBFF), the second from the low-surrogates range (\uDC00-\uDFFF).
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 30/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
BOM (Byte Order Mark) : BOM is a special Unicode character having code number of FEFFH, which is used to differentiate big-endian and
little-endian. For big-endian, BOM appears as FE FFH in the storage. For little-endian, BOM appears as FF FEH. Unicode reserves these two code
numbers to prevent it from crashing with another character.
UTF-8 file is always stored as big endian. BOM plays no part. However, in some systems (in particular Windows), a BOM is added as the first
character in the UTF-8 file as the signature to identity the file as UTF-8 encoded. The BOM character (FEFFH) is encoded in UTF-8 as EF BB BF.
Adding a BOM as the first character of the file is not recommended, as it may be incorrectly interpreted in other system. You can have a UTF-8 file
without BOM.
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 31/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
Worse still, there are also various chinese character sets, which is not compatible with Unicode:
GB2312/GBK: for simplified chinese characters. GB2312 uses 2 bytes for each chinese character. The most significant bit (MSB) of both bytes are
set to 1 to co-exist with 7-bit ASCII with the MSB of 0. There are about 6700 characters. GBK is an extension of GB2312, which include more
characters as well as traditional chinese characters.
BIG5: for traditional chinese characters BIG5 also uses 2 bytes for each chinese character. The most significant bit of both bytes are also set to
1. BIG5 is not compatible with GBK, i.e., the same code number is assigned to different character.
For example, the world is made more interesting with these many standards:
Notes for Windows' CMD Users : To display the chinese character correctly in CMD shell, you need to choose the correct codepage, e.g.,
65001 for UTF8, 936 for GB2312/GBK, 950 for Big5, 1201 for UCS-2BE, 1200 for UCS-2LE, 437 for the original DOS. You can use command "chcp" to
display the current code page and command "chcp codepage_number" to change the codepage. You also have to choose a font that can display
the characters (e.g., Courier New, Consolas or Lucida Console, NOT Raster font).
Hence, in sorting or comparison of strings, a so-called collating sequence (or collation) is often defined, which specifies the ranks for letters
(uppercase, lowercase), numbers, and special symbols. There are many collating sequences available. It is entirely up to you to choose a collating
sequence to meet your application's specific requirements. Some case-insensitive dictionary-order collating sequences have the same rank for same
uppercase and lowercase letters, i.e., 'A', 'a' ⇒ 'B', 'b' ⇒ ... ⇒ 'Z', 'z'. Some case-sensitive dictionary-order collating sequences put the
uppercase letter before its lowercase counterpart, i.e., 'A' ⇒'B' ⇒ 'C'... ⇒ 'a' ⇒ 'b' ⇒ 'c'.... Typically, space is ranked before digits '0' to
'9', followed by the alphabets.
Collating sequence is often language dependent, as different languages use different sets of characters (e.g., á, é, a, α) with their own orders.
Example: The following program encodes some Unicode texts in various encoding scheme, and display the Hex codes of the encoded byte
sequences.
import java.nio.ByteBuffer;
import java.nio.CharBuffer;
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 33/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
import java.nio.charset.Charset;
// Encode the Unicode UCS-2 characters into a byte sequence in this charset.
ByteBuffer bb = charset.encode(message);
while (bb.hasRemaining()) {
System.out.printf("%02X ", bb.get()); // Print hex code
}
System.out.println();
bb.rewind();
}
}
}
H i , 您 好 !
UTF-16LE: 48 00 69 00 2C 00 A8 60 7D 59 21 00 [2-4 bytes variable-length]
H i , 您 好 !
GBK: 48 69 2C C4 FA BA C3 21 [1-2 bytes variable-length]
H i , 您 好 !
Big5: 48 69 2C B1 7A A6 6E 21 [1-2 bytes variable-length]
H i , 您 好 !
Recall that in the UTF-16 encoding scheme, a BMP characters uses 2 bytes. It is the same as UCS-2. A supplementary character uses 4 bytes. and
requires a pair of 16-bit values, the first from the high-surrogates range, (\uD800-\uDBFF), the second from the low-surrogates range (\uDC00-
\uDFFF).
In Java, a String is a sequences of Unicode characters. Java, in fact, uses UTF-16 for String and StringBuffer. For BMP characters, they are the
same as UCS-2. For supplementary characters, each characters requires a pair of char values.
Java methods that accept a 16-bit char value does not support supplementary characters. Methods that accept a 32-bit int value support all
Unicode characters (in the lower 21 bits), including supplementary characters.
This is meant to be an academic discussion. I have yet to encounter the use of supplementary characters!
Let me know if you have a better choice, which is fast to launch, easy to use, can toggle between Hex and normal view, free, ....
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 35/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
The following Java program can be used to display hex code for Java Primitives (integer, character and floating-point):
In Eclipse, you can view the hex code for integer primitive Java variables in debug mode as follows: In debug perspective, "Variable" panel ⇒ Select
the "menu" (inverted triangle) ⇒ Java ⇒ Java Preferences... ⇒ Primitive Display Options ⇒ Check "Display hexadecimal values (byte, short, char, int,
long)".
Integer number 1, floating-point number 1.0 character symbol '1', and string "1" are totally different inside the computer memory. You need to
know the difference to write good and high-performance programs.
In 8-bit signed integer, integer number 1 is represented as 00000001B.
In 8-bit unsigned integer, integer number 1 is represented as 00000001B.
In 32-bit floating-point representation, number 1.0 is represented as 0 01111111 0000000 00000000 00000000B, i.e., S=0, E=127, F=0.
In 64-bit floating-point representation, number 1.0 is represented as 0 01111111111 0000 00000000 00000000 00000000 00000000
00000000 00000000B, i.e., S=0, E=1023, F=0.
In 8-bit Latin-1, the character symbol '1' is represented as 00110001B (or 31H).
If you "add" a 16-bit signed integer 1 and Latin-1 character '1' or a string "1", you could get a surprise.
Ans: (1) 42, 32810; (2) 42, -32726; (3) 0, 42; 128, 42; (4) 0, 42; -128, 42; (5) '*'; '耪'; (6) NUL, '*'; PAD, '*'.
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 37/38
2/25/2020 A Tutorial on Data Representation - Integers, Floating-point numbers, and characters
Feedback, comments, corrections, and errata can be sent to Chua Hock-Chuan ([email protected]) | HOME
https://www3.ntu.edu.sg/home/ehchua/programming/java/DataRepresentation.html 38/38