or one reader and several writers? and there are some extra output feature must be derived from open data first.
Page 1 / 1
What the best practice is depends on a lot of different things. Personally, I would use one workspace, one (Feature)Reader and multiple (Feature)Writers.