I was part of a project that knew this siren song well. Here's our Swift implementation:
Winternitz signatures are fascinating because they're secure while still being within the reach of most programmers and--as the article points out--they're quantum safe.
We never got around to implementing the Merkle tree portion because while it saved space it added more complication and still didn't quite solve the "one time" issue. Instead, we implemented a simple blockchain. Users would sign a block containing the signature of the content they wanted to sign and their next key. Again, within the capabilities of most coders to verify themselves, which was the goal.