PROJECT DOSSIER // [RUSH 2]VIEW SOURCEBACK TO PROJECTS →

01 // PROJECT AT A GLANCE

RUSH 2

C

Guessing the language of a text from its letter statistics alone.

A second weekend team rush: the program receives a text and must determine which language it is written in, with no dictionary. The idea is that every language has a statistical signature — its letter frequencies — stable enough to be recognised. The program therefore never proves its answer: it bets on the most likely one, which completely changes how you judge whether it is right.

Frequency analysis of a text's characters, comparison against reference language profiles and selection of the closest one.

PRIMARYC
DELIVERYTEAM RECORD
VERIFICATIONmake
SOURCEINSPECTOR READY
MEDIA ATTACHMENTS (Images, Photos, Videos)

02 // PROJECT SIGNALS

DATE:
2018
PRIMARY LANG:
C
GRADE:
B
DURATION:
1 afternoon
SECONDARY:
—
STACK:
Makefile / gcc

03 // KEY OUTCOMES

  • Language recognition with no dictionary, purely from statistics
  • The same principle as breaking a cipher by frequency analysis

BEYOND THE BRIEF

  • Three languages beyond the one required: French, German and Spanish, each in its own file
  • The irregularities of French (soixante-dix, quatre-vingts) and German (reversed order of units and tens) are handled

04 // SOURCE TREE

READING SOURCE…