A lot of data is unstructured, there are structured repositories too. Many of the unstructured repositories mostly live separately from the papers that are related to the data. Key point for this workshop, how can we bring ideas and evidence together again? A serious point is that there is a fear that most of the data in drug discovery cannot be reproduced, can have potential significant impacts on global health. What is a paper? - some tables, some figures, a lot of text. Providing the text as OA is fantastic, yet text mining has limitations. e.g. in extracting pathways. Figures are mostly data convolved into pixels in a way that makes it hard to extract that data. Figures are central to a formal scientific demonstration. They are part of the natural scientific workflow, they represent the structure of scientific investigation.