In addition to a "flag" link on submissions, add a "spam" link so that we can start differentiating flags for inappropriate content and outright spam. I would imagine the algorithm could trust spam flags by users a little more since it's almost universally understood when a submission is spam versus just controversial.