when displaying a hover card for a link "[ [ foo ] ]" , could it scan the target text and filter out "in bar, foo is blah blah blah" - and just display a compact definition. "... is a blah blah blah..." since the viewer already knows it's "foo" from the link , and the context ("in 'context' , foo is..") is probably not so far away.
Topic on Talk:Page Previews/Flow
This post was hidden by Quiddity (WMF) (history)