#! /bin/csh
# 
# This does a poor job of processing a set of manpage listed in
# the file "manpages". It finds the 40 most commonly occuring 
# words in the file, excluding some "stop words" given in the
# egrep commands.  The resulting file has a suffix .doc, and is
# sorted into alphabetical order - but has the word counts as
# well.
# 
#

foreach f (`cat manpages`)
echo $f
man $f \
| col -b \
| tr -cs 'A-Za-z' '\012' \
| tr 'A-Z' 'a-z' \
| egrep '^..*$' \
| egrep -v '^([a-z]|the|or|of|it|above|below|after|an|and)$'\
| egrep -v '^(on|not|that|used|use|see|then)$'\
| egrep -v '^(may|with|as|are|but|does|one|have|no)$'\
| egrep -v '^(in|if|to|from|is|be|by|use|do|like|only|you)$'\
| sort | uniq -c | sort -n | tail -40 | sort -k 2 > $f.doc

end