We consider a code construction techniques for the (d)-constraint
proposed by K. Immink in his book.  The techniques usually gives small
codes.  The technique was applied to phase change memory (PCM) by
K.~Cai in order to construct codes for a (k)-constraint with very
high efficiency.  If we use the Immink coding technique in designing
the code, we can easily adjust the code by appropriately choosing code
words so that a given performance index is optimized.  We interpret
the Immink coding technique in terms of the state splitting method and
characterize the coding technique using follower sets.