When numbers break their boundaries — and how to multiply without multiplying
On June 4th 1996, the Ariane 5 rocket exploded 37 seconds after launch, destroying a €370 million spacecraft. The cause? A 64-bit floating point number was converted into a 16-bit integer — the value was too large to fit, overflow occurred, and the navigation system crashed. The entire mission was lost to a bug that took engineers minutes to find but cost a decade of development. Meanwhile, binary shifting is used by every image processor, audio encoder, and graphics engine on the planet — shifting bits left multiplies values, shifting right divides them, without a single multiplication instruction. Faster. Cheaper. Elegant. These two topics sit on either side of the same truth: bit boundaries are a hard physical limit, and every professional programmer must know what happens at the edge.
Overflow occurs in binary addition when the result is too large to be stored in the number of bits available. In an 8-bit system, the maximum value is 255. Any result above 255 cannot be correctly stored — a carry bit is generated out of the Most Significant Bit (MSB) column, and that bit is lost. The stored result becomes incorrect.
Perform the addition normally. Write the carry row clearly above the numbers.
Check the carry out of the leftmost (128s / MSB) column. If there is a carry out → overflow has occurred.
Quick denary check: add the two denary values. If the sum exceeds 255 → overflow will always occur in 8 bits.
State the overflow explicitly in your answer. Write: "Overflow has occurred — the result exceeds 255 and cannot be stored correctly in 8 bits."
Confusing overflow with a wrong answer — overflow is a specific condition: carry out of the MSB. A wrong carry in the middle of the sum is just an arithmetic error, not overflow.
Not stating overflow explicitly — even if your addition is correct, you must write "overflow has occurred" to earn the dedicated mark.
Writing the 9-bit result — in an 8-bit system, the overflow carry bit is lost. The stored answer is only the 8 rightmost bits, even though they are wrong.
Work through each step before revealing — this is a classic 4-mark exam question.
Show the carry row in every addition question.
Binary shifting moves all bits in a number a set number of positions left or right. It is the most efficient way a CPU can multiply or divide by powers of 2 — much faster than performing actual multiplication. Vacated positions are filled with 0s. Bits that fall off the end are lost.
Filling with 1s instead of 0s — vacated positions are always filled with 0s in a logical shift. Never 1s.
Confusing the direction of multiplication — left shift = bigger number (×2). Right shift = smaller number (÷2). Students sometimes reverse this.
Not accounting for lost bits — if a 1 bit is shifted off the left edge during a left shift, the result will be wrong (overflow). Always check what was lost.
A classic exam-style question combining the shift operation with the multiplication equivalence.
Show the full bit layout — before and after — in your answer.
3 overflow questions + 3 shifting questions. Answer all 6 for your XP badge.