The strain on scientific publishing