Towards Legally Enforceable Hate Speech Detection for Public Forums