Lempel-Ziv-Welch algorithm in ruby

Here is a small script I wrote for the fun of it. Originally I wanted to put it on twitter, but it got a bit too large for it. This made the code very compact, but also quite unreadable. However, I just made a blog post on it, and don't plan to expand or modify it any further. Pretty simple being, it converts a string to a Lempel-Ziv-Welch (LZW) encoded (semi-human-readable) string. Enjoy.

# Variable initialization
s = 'to be or not to be. ' * 100
t = ''
d = ['']
c = ''
l = false

# Algorithm
s.each_byte do |b|
 l = true
 t << b.chr
 unless d.include? t 
  d.push t
  c << d.index(t.chop).to_s << t[t.length-1].chr
  t = ''
  l = false
 end
end

if l == true
 c << d.index(s[s.length-1].chr).to_s
end

# Output
puts 'original: ' << s.length.to_s
puts 'compressed: ' << c.length.to_s
p c

The benefit of this 'compression' isn't high, but it's something. For example, "to be or not to be. " is converted from 20 to 27 "compressed" characters. However, the same phrase, but used 100 times in succession (as in the example) gets compressed from 2000 to 846 characters.