I first thought using empty regular expression ('') as FS, but it
doesn't work.
I have this requirement because I am working on an awk script that
analysis Chinese ideograph in text files. I want to process routine for
each single ideograph. I guess this is a wired requirement for western
text because they are usually counted by word and separated by space
(where Chinese ideograph form words that are not separated by anything,
it happens a whole Chinese novel doesn't contain a single space.)