word_differences.sh

#!/bin/sh

# Show the differences between two Microsoft Word files.

echo Detecting differences between $1 and $2.
echo MD5 hashes:
md5 "$1" "$2"

# Unpack the contents of the documents. The Word .docx format is a ZIP file containing XML files.
mkdir "$1".contents
mkdir "$2".contents
unzip -d "$1".contents "$1"
unzip -d "$2".contents "$2"

# Run an XML formatter on each XML file in the document.
find "$1".contents -name '*.xml' -exec xmllint --format --output {}.formatted {} \; -exec mv {}.formatted {} \;
find "$2".contents -name '*.xml' -exec xmllint --format --output {}.formatted {} \; -exec mv {}.formatted {} \;

# Recursively find the differences.
diff --unified --recursive "$1".contents "$2".contents > differences.txt