No need for something like that. The Trustnet paper already provides human-meaningful labels. Brugeman, any plans to implement those labels to make ranking easier?