Reinvent match and rewrite so that they handle sequences of
words, not just single words. But make sure they still do
work for single words; don't lose any functionality.

Examples

> (setq A1 (match '(the 1 sat on the 2 quietly)
                  '(the big fluffy cat sat on the bad dog quietly)))

> A1
((2 (BAD DOG)) (1 (BIG FLUFFY CAT)))

> (rewrite A1 '(quiet sitter = 1 -- sittee = 2))
(QUIET SITTER = BIG FLUFFY CAT -- SITTEE = BAD DOG)



> (setq A2 (match '(1 says I 2)
                  '(Lisa says I look like a parrot)))
> A2
((2 (LOOK LIKE A PARROT)) (1 (LISA)))

> (rewrite A2 '(how does it make you feel about 1 when 1 claims you 2 ?))
(HOW DOES IT MAKE YOU FEEL ABOUT LISA WHEN LISA CLAIMS YOU LOOK LIKE A PARROT ?)



> (setq A3 (match '(1 says I 2)
                  '(my aunt says that I should learn to knit)))
> A3
FAIL


It is OK if your match can not handle situations like this
(match '(the 1 sat on the 2)
       '(the man who sat on my dog also sat on the policeman's dog))
even though
       1 = (man who sat on my dog also)
       2 = (policeman 's dog)
is the correct solution.
But you should put some thought into how you could make it work properly
in cases like that.