Experiments with five letter words.
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
output
.gitignore
LICENSE
README.md
diff_by_n.py
diff_by_one.py
distinct.py
euclidean_distance.py
get_words.py
letter_coverage.py
lexico.py
near_palindromes.py
palindromes.py
reverse_lexico.py
sgb-words.txt
stats.py

README.md

five-letter-words

This repository contains Donald Knuth's GraphBase list of five-letter words, as well as scripts to run various combinatoric experiments, graph algorithms, and other algorithms to explore the relationships among these words.

The list of words comes from [1] and is in the public domain.

Get Words

A Python program that contains a method for getting all of the five letter words from a file, and that's about it.

Warm Up Exercises

Exercises 26-37 of Knuth's Volume 4 Fascile 0 are intended as a warm up to get to know the SGB five letter word list. Solutions to these exercises are listed below.

distinct.py- computes the number of SGB words containing exactly k distinct letters.

diff_by_one.py - computes the number of words in the SGB that are off by a single letter, shifted a single place. For example, "might" and "night" or "large" and "marge". There is a surprisingly large number of such pairs.

euclidean_distance.py - computes the euclidean distance between two words. This uses the traditional Euclidean distance definition but reinterprets distance to mean edit distance.

lexico.py - find words that are sorted by lexicographic order (front to back, a-z).

palindromes.py - look for five letter words that are either a palindrome, or a palindrome pair.

Variations

diff_by_n.py - computes words in SGB that have an edit distnace of n.

reverse_lexico.py - variation on lexico.py that finds words whose letters are in reverse lexicographic order.

Letter Coverage

letter_coverage.py - computes coverage of the alphabet (minimum number of words required to provide X letters of the alphabet)

Knuth mentions, in the text, a couple of facts about how many words cover how much of the alphabet. We authored a dynamic program to compute precisely this - given a number of letters N from the alphabet, this program computes the minimum number of words it takes to cover all N letters.

Also see https://charlesreid1.com/wiki/Letter_Coverage..

Sources

  1. Knuth, Donald. The Stanford GraphBase: A Platform for Combinatorial Computing. New York: ACM Press, 1994. <http://www-cs-faculty.stanford.edu/~knuth/sgb.html>