In this brief note we make a simple observation which relates the maximum matching problem to the family of vertex partition problems, many members of which are known to be NP-complete. Based on this observation we suggest a large number of unsolved problems, algorithmic solutions to any one of which would generalize standard matching algorithms, yet could still run in polynomial time.