Problems
Number bases and coding
Binary, octal, hexadecimal and the coding schemes built on top of them: the same number, written another way.
21 problems · Showing 1–15 of 21
0x2A in decimal — the answer to life, the universe and everything
A joke from 1979 that doubles as a clean base-conversion exercise: two hexadecimal digits, one decimal answer.
Level: Classic result · 2026-11-09
Why −1 shows up as FFFF
Enter minus one as decimal and the base converter's 16-bit row answers with four F's — the machine-word result every embedded interview asks about.
Level: Exam and contest · 2026-11-09
2³¹ − 1 is 2147483647, and it is prime
The largest signed 32-bit integer, 31 ones in binary, and the largest known prime for over a century.
Level: Classic result · 2026-11-10
Why chmod 755 is really 493
A Unix file mode is one number written in base 8 — three permission triples, one digit each.
Level: Classic result · 2026-11-10
What number is #FFFFFF?
Twenty-four bits of ones, 16777215 in decimal, and one fewer than 256 cubed.
Level: Classroom · 2026-11-10
0xDEADBEEF, and the number it hides
A debugging sentinel people recognise, and the arithmetic that puts it outside a signed 32-bit integer.
Level: Classic result · 2026-11-10
Machine epsilon: the smallest change a computer admits at 1
2⁻⁵² is 2.2204460492503⋅10−16, and adding it to 1 does not move the printed number at all.
Level: Classic result · 2026-11-10
0.1 + 0.2 does not equal 0.3 — but the screen says it does
The most famous floating-point bug report, and the subtraction that makes the residue visible.
Level: Classic result · 2026-11-10
At 2^53 the screen stops being able to count by one
9.007199254741⋅10+15 is the last integer with a neighbour, and a subtraction shows the gap.
Level: Classic result · 2026-11-10
A Babylonian tablet gives √2 as 1;24,51,10 in base 60
YBC 7289, ca. 1800–1600 BCE: the diagonal of a unit square, correct to six decimals, four thousand years early.
Level: Classic result · 2026-11-10
IPv4 has 2³² addresses, and the workbench will not print all of them
4,294,967,296 is what 32 bits buys — and where floating point starts hiding digits behind an exponent.
Level: Classic result · 2026-11-10
A UUID v4 has 5.3169119831397⋅10⁺³⁶ variants, and 2⁶⁴ − 1 is where doubles run out of digits
122 random bits once the version and variant fields are carved out — plus an exact 2⁶⁴ − 1 in the base converter and a rounded one on the workbench.
Level: Classic result · 2026-11-10
log₂ 1024 is exactly 10, and the change-of-base quotient says so
log(1024)/log(2) prints a whole 10; 2^10 is 1024, and the binary form is a 1 followed by ten zeros.
Level: Classroom · 2026-11-10
255 is eight bits full: FF in hex, 11111111 in binary, 2^8 − 1
The ceiling of an unsigned byte appears as 255 in colour channels, IPv4 octets and checksums alike.
Level: Classic result · 2026-11-10
Fermat's fifth number is 4294967297, and Euler showed it is not prime
Two multiplications land on the same ten-digit mantissa, and the hexadecimal screen says why the number looked special.
Level: Classic result · 2026-11-10