Did you know that the eight bits 11111111 can be interpreted as 255, −1 or a character-code value, depending on the representation the question specifies? The bits themselves do not tell us which interpretation is intended. Computers store patterns; people and programs supply the rules that make those patterns meaningful.
Ministry of Education V3.0 Reference 358 explains the 2027 Singapore–Cambridge SEC G3 Computing K349 syllabus area 1.2, Representation of Data. Students learn to read binary place values, convert positive whole numbers between binary, denary and hexadecimal, use fixed-width two’s complement for negative integers, and understand text encoding through the syllabus example of eight-bit extended ASCII. The guide includes illustrated arithmetic in words, a twelve-stage checking routine, twenty-four original worked cases, common-error diagnostics and practice questions.
This independent eduKateSG educational resource is not an official MOE, SEAB or Cambridge publication. It follows the publicly available 2027 K349 subject specification, but teachers and candidates should consult the official document for examination scope and administration. All computing scenarios and questions below are original teaching examples, not actual SEC examination questions. We use an eight-bit word in many exercises to make the underlying reasoning visible; when the question specifies a different width, use that width instead.
Study route: learn the meaning before memorising the method
Start with the section on place value and write one binary-to-denary conversion by hand. Next convert the same number to hexadecimal, then reverse the process to verify it. Only after positive-number conversions are stable should you begin signed two’s complement. Finish with text encoding, because a character stored in a byte is not necessarily the number displayed by the same byte.
A useful first diagnostic is to explain what 10110110 means in both unsigned eight-bit and signed eight-bit two’s complement. It is 182 if interpreted as unsigned. It is −74 if interpreted as signed two’s complement. Both readings are possible, but never interchangeable without a rule. Every conversion exercise should begin by naming the representation.
1. The official 2027 K349 learning boundary
Section 1.2 asks students to explain the use of binary to represent positive whole numbers, to convert positive whole numbers between binary, denary and hexadecimal, to use two’s complement for positive and negative whole numbers in a fixed number of bits, and to explain how information can be represented in bits using an eight-bit extended ASCII encoding for English text as an example.
That is a focused syllabus boundary. The broader subject also covers hardware, networks, algorithms, programming and logic, but the task here is data representation. Do not let a question about a byte’s numeric interpretation drift into advanced floating-point standards or arbitrary Unicode implementation details that are beyond the stated examples. Understanding the limits of the particular representation is more useful than collecting unrelated technical facts.
2. A bit is a binary digit, not a tiny decimal digit
A bit takes one of two values, 0 or 1. A series of bits may encode a number, a text character, a colour value or part of an instruction, depending on the system using it. Eight bits form a byte. The same bit sequence can be interpreted differently because representation is a convention, not a visible label physically attached to the bits.
For numbers in ordinary positional binary, the place values from right to left are 1, 2, 4, 8, 16, 32, 64, 128 and so on. Each place is a power of two, just as denary place values use powers of ten. Leading zeros do not change an unsigned integer’s numerical value, but they can matter when a question specifies an eight-bit width.
3. Convert binary to denary by summing the active positions
Take 10101101 in binary. Reading its eight positions from left to right gives values 128, 64, 32, 16, 8, 4, 2 and 1. Ones occur in the 128, 32, 8, 4 and 1 positions. The total is 128 + 32 + 8 + 4 + 1 = 173. That is its unsigned denary value.
Checking the result is an important habit. Since the highest active place is 128 and every other active contribution is positive, the answer must be between 128 and 255. If you instead obtain a number below 128, your place values or arithmetic must be wrong. For another quick check, 10000000 is 128 unsigned, and 11111111 is 255 unsigned. These landmarks catch many conversion mistakes.
4. Convert denary to binary by subtracting powers of two
To convert 45 to an eight-bit binary representation, choose the largest suitable place value, 32. The remainder is 13. Choose 8, leaving 5. Choose 4, leaving 1. Choose 1, leaving 0. The active places are 32, 8, 4 and 1. Their eight-bit pattern is 00101101. Checking gives 32 + 8 + 4 + 1 = 45.
Repeated division by two is another valid technique. Divide by two and record the remainder each time; read the remainders in reverse order. Both methods work, but neither should be used mechanically without checking that the output matches the original denary quantity. Students who remember the method yet reverse the remainder order can obtain an incorrect pattern that looks plausible.
5. Hexadecimal is a compact notation for groups of four bits
Hexadecimal is base sixteen. Its digits are 0 through 9, followed by A, B, C, D, E and F representing ten through fifteen. Every hexadecimal digit maps to exactly four binary bits. A binary nibble 1010 corresponds to A; 1111 corresponds to F. This makes hexadecimal convenient for describing long binary patterns.
For example, split 11110000 into groups 1111 and 0000, giving F0 in hexadecimal. The value is 15 × 16 + 0 = 240 in denary. A leading notation such as 0xF0 may be used to indicate hexadecimal, but the digits F0 alone need a stated base when ambiguity is possible. Never read hexadecimal as though it were denary merely because some digits look familiar.
6. Hexadecimal to binary and back should be reversible
Consider 0x2F. The digit 2 corresponds to 0010 in binary; F corresponds to 1111. Together they give 00101111. In denary, the value is 2 × 16 + 15 = 47. The independent binary check yields 32 + 8 + 4 + 2 + 1 = 47. Both paths agree.
Reverse conversions are especially useful when learning hexadecimal. Given 10111010, group it as 1011 1010, giving BA in hexadecimal. The denary value is 11 × 16 + 10 = 186. A mistaken grouping from the wrong end can change every hexadecimal digit; group from the right and add leading zeros only if necessary for a complete four-bit group.
7. Fixed width changes what patterns can represent
An eight-bit word has exactly 2 to the power of 8, or 256, distinct patterns. If all are interpreted as unsigned integers, they represent values from 0 through 255. The binary sequence 10000000 is therefore 128 unsigned. It is not automatically negative merely because its first bit is 1.
When a number uses signed two’s complement, the same eight-bit patterns instead cover −128 through +127. The high-order bit is involved in distinguishing negative values, but a safe explanation of two’s complement must state the bit width and interpretation. Values outside the signed range cannot be represented correctly using that width.
8. Two’s complement is a signed number convention
For a negative number in fixed-width two’s complement, begin with the positive magnitude in the specified number of bits, invert every bit, and add one. For −13 in eight bits: +13 is 00001101; bitwise inversion gives 11110010; adding one gives 11110011. Under eight-bit two’s complement that final pattern represents −13.
The method works for any negative value representable in the chosen width, with special care at the minimum value. To interpret a negative pattern, invert all eight bits and add one to find its magnitude, then attach the minus sign. For 11000110, inversion gives 00111001; adding one gives 00111010, which is 58. Therefore 11000110 represents −58 as signed eight-bit two’s complement, while its unsigned interpretation is 198.
9. The endpoints help reveal mistaken answers
The eight-bit signed range −128 to +127 is not symmetrical. There is one more negative integer because zero occupies a pattern and there are 256 patterns altogether. The all-zero pattern 00000000 represents zero; there is no separate negative zero. The pattern 01111111 is +127 and 10000000 is −128.
The pattern 11111111 means −1 under signed eight-bit two’s complement but 255 under unsigned interpretation. Students sometimes attempt to represent +128 in signed eight bits as 10000000, which actually means −128. The error is not in the arithmetic for 128; it is the failure to check whether 128 lies inside the signed representable range.
10. Signed and unsigned interpretations are not synonyms
A byte such as 10110110 has the unsigned value 128 + 32 + 16 + 4 + 2 = 182. Its signed value in eight-bit two’s complement is 182 − 256 = −74. This subtraction shortcut works for interpreting the negative half of the eight-bit signed range because subtracting 256 accounts for the wraparound relationship.
Try the opposite case: 01011010 is 90 unsigned and 90 signed because its most significant bit is zero. When the highest bit is zero, signed and unsigned non-negative values agree. When the highest bit is one, the interpretations diverge. This is not a contradiction; the data were never accompanied by a unique intended meaning.
11. Addition reveals why width and overflow matter
A computer can add fixed-width binary patterns using ordinary binary carries. In eight-bit arithmetic, +7 is 00000111 and −3 is 11111101. Adding them gives a nine-bit working result of 1 00000100. Keeping the lower eight bits leaves 00000100, representing +4. The discarded carry is part of fixed-width arithmetic, not proof that the mathematical result is wrong.
Now consider signed eight-bit +100 plus +50. The correct mathematical total is +150, outside the signed range. The bitwise eight-bit result is 10010110, which a signed eight-bit reader interprets as −106. This is signed overflow, not an actual conclusion that adding positive numbers creates a negative quantity. State both the intended mathematical result and the representational limitation.
12. A binary carry is different from signed overflow
In fixed-width arithmetic, a carry beyond the highest retained bit is not by itself a reliable test of signed overflow. Two positive representable numbers can add to a result whose sign bit becomes one even without a ninth carry. Conversely, adding a positive and a negative number may produce a carry beyond the width while leaving a valid signed result.
Students should reason about the expected mathematical answer and the signed range rather than rely on a single visual indicator. An arithmetic expression may have a correct mathematical answer that the chosen encoding cannot contain. Showing this distinction demonstrates conceptual understanding of representation, not just competence with binary addition.
13. Text and numbers can use the same bits differently
The K349 syllabus uses an eight-bit extended ASCII encoding of English text as an example of how bit patterns can represent information. In standard ASCII, capital A has code value 65, which is 01000001 when written in an eight-bit container; lowercase a is code value 97, or 01100001. The bits can be read as numbers, but in a text context the code values identify characters.
Be precise about the word ‘extended’. Standard ASCII is a seven-bit encoding with code values 0 through 127. Different eight-bit extended ASCII code pages disagree about some values from 128 through 255. Therefore do not assume that every possible eight-bit value names the same character across all systems. For school exercises, follow the particular character table and rules provided.
14. An eight-bit encoding cannot describe every modern character
One byte contains 256 possible bit patterns. That is sufficient for many basic English text examples but not for all the world’s writing systems and symbols. Modern text commonly uses Unicode with an encoding such as UTF-8, where different characters may require different numbers of bytes. Unicode is a wider context, not a reason to ignore the specified K349 extended ASCII example.
If a question supplies a table giving a character’s eight-bit code, use that table. Do not invent the code of an unfamiliar symbol or claim every character occupies one byte. Characters, numeric values and storage units should be kept conceptually distinct. A student can memorise 65 for ‘A’ and still misunderstand encoding if they cannot explain why a byte needs a shared interpretation.
15. Bits, bytes and file-size labels require the right unit
Eight bits make one byte. Thus 1,000 bytes contain 8,000 bits, before considering any additional coding or transmission overhead. The prefix kilo in SI usage means 1,000, while kibi means 1,024; 1 kB is 1,000 bytes and 1 KiB is 1,024 bytes. A question’s stated convention should be followed exactly.
Storage figures must not be confused with the meaning of the information stored. A text file’s size depends on the number of encoded bytes plus any structural information, not simply the number of visible words. A displayed numeral such as ‘123’ is three characters in a simple single-byte encoding, whereas a binary integer can represent the numeric value 123 using eight bits. These are different representations of different kinds of content.
16. Good conversion checks are independent of the original route
After converting 0xAF to binary 10101111, verify its value in denary as 10 × 16 + 15 = 175, then independently sum the binary active positions: 128 + 32 + 8 + 4 + 2 + 1 = 175. Agreement increases confidence. It does not guarantee that the initial source was copied correctly, so review the question as well.
When converting signed values, always test the result in both directions. For −5 in eight bits: +5 is 00000101, inversion is 11111010, and adding one produces 11111011. Reversing that bitwise process returns +5 as a magnitude. The ability to move in both directions is more resilient than recalling one rehearsed example.
17. A twelve-stage Read–Convert–Verify protocol
This teaching protocol is not an official examination marking scheme. It creates a consistent workflow for calculating, annotating and checking unfamiliar representation questions. The same habit scales from a few bits to larger systems: clarify what is represented, apply the correct rule, respect the width and verify with a different route.
Stage 1 — Identify the intended quantity
Ask whether the problem describes an integer, a bit pattern, a character or a storage amount. Do not choose the calculation method from the digits alone.
Stage 2 — Name the original base
Check whether an input is binary, denary or hexadecimal. A sequence of decimal-looking digits can still be hexadecimal when its base is stated.
Stage 3 — Name the target form
State what the answer must look like. A binary answer with the right denary value may still miss an explicitly required hexadecimal notation.
Stage 4 — Check bit width
Record whether the question requires eight bits, four bits or another fixed width. Insert leading zeros where the representation permits and the width requires.
Stage 5 — Check signedness
Determine whether the representation is unsigned or signed two’s complement. If no interpretation is supplied, say what assumption you have made.
Stage 6 — Mark place values
For positional conversion, write the relevant powers of two or sixteen. Label positions rather than trying to recognise a long pattern by sight.
Stage 7 — Show one transparent method
Use repeated division, powers-of-two subtraction, a weighted sum or four-bit grouping. A short trace makes errors diagnosable.
Stage 8 — Handle negative values deliberately
For representable negatives in two’s complement, invert the fixed-width magnitude bits and add one. Keep every bit position visible.
Stage 9 — Test the valid range
Before accepting an answer, check the unsigned or signed range of the given width. Arithmetic may be correct mathematically but unrepresentable.
Stage 10 — Reverse the conversion
Convert the proposed answer back to the original quantity by a different route when possible. For hexadecimal, a four-bit check is often fast.
Stage 11 — Apply the correct interpretation
If the question is about text, use its character encoding table; if about integer arithmetic, use signedness and width; if about capacity, use units.
Stage 12 — Explain any limitation
State whether overflow, insufficient width, ambiguous encoding or missing context affects the answer. Such a statement is an honest part of computing literacy.
18. Twenty-four worked conversion and interpretation cases
These original worked examples practise what often goes wrong: place-value reversal, hexadecimal grouping, confusing signed and unsigned values, mishandling negative magnitudes and assuming a text character is a number. Use the cases as models of reasoning, not answers to copy without understanding. Whenever possible, cover the worked answer and attempt the case independently first.
Worked case 1 — Binary 10101 to denary
The active positions are 16, 4 and 1. Their sum is 21. If an eight-bit answer is required, the same unsigned value may be written 00010101. The extra leading zeros do not contribute a quantity.
Worked case 2 — Binary 10101101 to denary
The active positions are 128, 32, 8, 4 and 1. The sum is 173. A student who reads 10101101 as a denary quantity has skipped the instruction to interpret its base.
Worked case 3 — Denary 26 to eight-bit binary
26 = 16 + 8 + 2, so the active places are 16, 8 and 2. The result is 00011010. Read the answer back as 16 + 8 + 2 to confirm it.
Worked case 4 — Denary 45 to binary and hex
45 = 32 + 8 + 4 + 1. In eight bits this is 00101101; grouping 0010 1101 gives hexadecimal 2D. As a second check, 2 × 16 + 13 = 45.
Worked case 5 — Denary 240 in binary and hex
240 = 128 + 64 + 32 + 16, producing 11110000. The two four-bit groups are 1111 and 0000, giving F0. The hexadecimal check is 15 × 16 = 240.
Worked case 6 — Hexadecimal 0x3C to denary
The digits are 3 and C, where C means 12. The value is 3 × 16 + 12 = 60. In binary the two four-bit groups are 0011 and 1100, making 00111100.
Worked case 7 — Hexadecimal 0x2F to binary
2 becomes 0010 and F becomes 1111. Joined, the result is 00101111. As a denary check, 32 + 8 + 4 + 2 + 1 = 47.
Worked case 8 — Binary 10111010 to hexadecimal
Group from the right: 1011 1010. The first group is B and the second A, so the result is BA. The denary check gives 11 × 16 + 10 = 186.
Worked case 9 — Hexadecimal 0xAF to denary
A represents 10 and F represents 15, producing 10 × 16 + 15 = 175. Binary 1010 1111 also confirms 128 + 32 + 8 + 4 + 2 + 1 = 175.
Worked case 10 — The same eight bits interpreted differently
The pattern 10110110 is 182 if unsigned. Under signed eight-bit two’s complement it is 182 − 256 = −74. Naming both interpretations without the distinction is not a valid single answer.
Worked case 11 — Interpret 11001010 as signed eight-bit
The unsigned value is 128 + 64 + 8 + 2 = 202. Since the sign bit is one, signed interpretation is 202 − 256 = −54. The equivalent invert-and-add-one magnitude check gives 54.
Worked case 12 — Represent −13 using eight-bit two’s complement
Write +13 as 00001101. Invert to 11110010, then add one to reach 11110011. Its unsigned numerical value is 243, and 243 − 256 = −13, verifying the signed result.
Worked case 13 — Represent −58 using eight-bit two’s complement
The positive magnitude is 00111010. Invert every bit to obtain 11000101, then add one to obtain 11000110. The unsigned reading is 198, which maps to −58 when interpreted as signed.
Worked case 14 — Interpret all ones
11111111 is 255 in unsigned eight-bit representation. In eight-bit signed two’s complement it is −1. If a question merely supplies the pattern without context, it has not yet specified one unique integer.
Worked case 15 — Interpret the signed minimum
Eight-bit 10000000 is −128 under two’s complement. There is no corresponding positive +128 in the signed eight-bit range. Do not attempt to force an out-of-range value into the encoding.
Worked case 16 — Check the signed maximum
01111111 is +127 in both eight-bit unsigned and signed conventions. Adding 1 mathematically gives 128, but the eight-bit signed result 10000000 denotes −128: signed overflow has occurred.
Worked case 17 — Encode −5 and recover the value
Positive 5 is 00000101. Invert to 11111010 and add one to obtain 11111011. To decode, invert 11111011 to 00000100 and add one, giving magnitude 5. The decoded signed value is −5.
Worked case 18 — Add +7 and −3 in eight bits
Use 00000111 + 11111101. The full working sum has an outgoing carry; the retained eight bits are 00000100, or +4. No signed overflow occurs because the mathematical answer fits the range.
Worked case 19 — Detect signed overflow when adding positives
100 is 01100100, and 50 is 00110010. Their eight-bit sum is 10010110. The true answer is 150, beyond signed eight-bit maximum 127; the stored pattern reads as −106 if signed.
Worked case 20 — Add two negative values
−5 is 11111011 and −4 is 11111100. The eight-bit arithmetic result is 11110111 after discarding a carry beyond the width. It is −9 signed, consistent with the mathematical sum.
Worked case 21 — Capital A in an ASCII example
Standard ASCII assigns capital A code 65. Writing that code with eight bits gives 01000001. In a text context it is the letter A, even though a numeric reader would interpret the same bits as 65.
Worked case 22 — Uppercase versus lowercase character code
The standard ASCII code for lowercase a is 97, or 01100001 when padded to eight bits. Capital A is 01000001. Text encoding distinguishes these characters rather than treating them as typographic presentations of one code.
Worked case 23 — How many values fit into eight bits?
Two choices at each of eight positions create 2^8 = 256 distinct combinations. The unsigned range is 0 through 255; signed two’s complement uses the same total of 256 patterns for −128 through +127.
Worked case 24 — File size and transmission quantities
A file of exactly 1,000 bytes contains 8,000 bits of stored payload. This is 1 kB in decimal SI storage terminology, but less than 1 KiB. The bit count does not automatically include networking overhead.
19. Diagnostic exercise: find the mistaken assumption
Mistake A: a student says that hexadecimal C equals 13. Correct it: C equals 12; D equals 13. Mistake B: a student groups the binary digits 111001 from the left as 1110 01. Group from the right and pad on the left: 0011 1001 = 0x39. Mistake C: a student says 10000000 means +128 in signed eight-bit two’s complement. Correct it to −128 and identify the range restriction.
Mistake D: a candidate calculates that 100 + 50 equals −106. The mathematical result is 150; the apparent −106 comes from interpreting an overflowing eight-bit pattern as signed. Mistake E: a student states that 0xF0 equals 150 because it contains F and zero. Base sixteen gives 15 × 16 + 0 = 240. Mistake F: someone assumes every extended ASCII system maps the high values to the same character. The high-half mapping can vary by code page.
20. Quick retrieval practice with checked answers
Question one: write 19 in eight-bit binary. Answer: 00010011, because 16 + 2 + 1 = 19. Question two: convert binary 01101000 to hexadecimal and denary. Answer: 0x68 and 104, since 64 + 32 + 8 = 104. Question three: convert 0xB4 to unsigned denary. Answer: 11 × 16 + 4 = 180. Question four: is +130 representable in signed eight bits? No; the positive limit is 127.
Question five: represent −2 in signed eight-bit two’s complement. Answer: 11111110. Question six: interpret 11111110 signed and unsigned. Answer: −2 and 254 respectively. Question seven: is 0x41 the same as the English letter A? Only when a specified text encoding maps code value 65 to A; in a numeric context it is the number 65. Question eight: how many bits are in three bytes? Exactly 24 bits.
21. Building a reliable student workflow
For a five-minute study session, take one random unsigned eight-bit pattern and convert it to denary and hexadecimal. For the next five minutes, use the same pattern as signed eight-bit two’s complement. Then write the comparison in a complete sentence: ‘These bits are … unsigned but … signed because the interpretation changes.’ The verbal explanation exposes misunderstandings that a calculator can conceal.
Tutors can ask students to show one intermediate place-value check and one inverse calculation, rather than merely inspecting a final answer. Parents can help with a simple flash-card exercise distinguishing a base from a sign convention. Neither needs specialist equipment. The important progress indicator is that the learner explains why the method is valid and notices when a requested value cannot fit.
22. Frequently asked questions
These short answers address recurring sources of confusion in binary and hexadecimal learning. The official K349 syllabus remains the reference for scope; the examples here provide practice in reasoning, including uncertainty when a representation has not been specified.
Why does a computer use binary?
Digital systems can represent two distinguishable logical states reliably, which makes binary a convenient basis for encoding information and processing it. Binary is an information representation, not a claim that every physical computer component is literally a small written zero or one.
Does a leading zero change a binary number?
For an unsigned positional integer, leading zeros do not change the value. They may be important when a question requires a fixed bit width or when discussing the signed interpretation of a particular byte.
Is hexadecimal bigger than binary?
Hexadecimal is a different positional notation using base sixteen. A hexadecimal digit encodes the same information as four binary digits, so it is more compact to write, not inherently a larger numerical value.
What does the F in hexadecimal mean?
It is one digit with numerical value fifteen. The two-character hexadecimal number F0 equals 15 × 16 + 0 = 240 denary. The meaning depends on the base, not the appearance of an alphabetic character.
Why is negative zero missing in two’s complement?
The all-zero pattern represents zero. Two’s complement uses its remaining bit patterns to cover negative and positive integers, yielding one more negative value than positive value at a fixed width.
Can eight signed bits store +128?
No. In eight-bit two’s complement, the maximum positive value is +127. A bit pattern of 10000000 represents −128 signed. A wider signed representation is required for +128.
Is a carry the same thing as overflow?
No. A carry outside a fixed number of bits and signed overflow are distinct phenomena. Signed overflow occurs when the true mathematical result is outside the representable signed range. The final sign and operands must be interpreted together.
Is ASCII always eight bits?
Standard ASCII uses seven bits for code values 0 through 127. The syllabus uses the example of an eight-bit extended ASCII encoding; mappings beyond the standard range can differ among code pages.
Can one byte represent every written character?
No. A single byte has only 256 distinct bit patterns. Modern Unicode text encodings, such as UTF-8, can represent far more characters and may use multiple bytes for a character.
What is the difference between kB and KiB?
One kB is 1,000 bytes in SI decimal usage; one KiB is 1,024 bytes. In a calculation, read the label and the convention stated by the question rather than assuming every use of ‘kilobyte’ means the same thing.
Which is more useful: memorising conversions or understanding them?
Memorising small values speeds up common problems, but understanding place values, grouping, sign and width is essential for checking unfamiliar patterns. The goal is to justify the answer rather than recognise only rehearsed strings.
What should I do when the question leaves signedness unspecified?
Point out the ambiguity and, if an answer is required, state your assumption. For example, the byte 11111111 is 255 unsigned and −1 under signed eight-bit two’s complement; there is no unique integer interpretation without context.
Official 2027 syllabus and connected eduKateSG references
Consult the official 2027 SEC G3 Computing K349 syllabus and the official 2027 G3 syllabus directory. Continue to Reference 321 — Computing Logic Gates, Reference 342 — Computing Manual Algorithms, the MOE V3.0 reference library and the immutable Clementi instructional standard.
Editorial review: 29 October 2026, Singapore. The worked cases are original classroom examples. The underlying principle is simple: do not ask what a pattern means until you know the base, width, sign rule and purpose of the representation. A correct answer describes both the value and the conditions under which that answer is valid.
MOE V3.0 sequence navigation
Previous: Reference 357 — Nutrition and Food Science Coursework · MOE V3.0 library.