well, how do you define a spam? Is it any document that has been constructed around some key phrases or some urls that are in the spam database? If that's the

well, how do you define a spam? Is it any document that has been constructed around some key phrases or some urls that are in the spam database? If that's the case, why not just do a similar document evaluation see if the similarity of the document and some sample in the DB reaches certain threshold?