While I agree that sharing and opening up data etc. is important (not only to profit from it as start ups, but primarily for science itself), I don't think this

While I agree that sharing and opening up data etc. is important (not only to profit from it as start ups, but primarily for science itself), I don't think this 'one library' idea is practically possible. Different fields use different categorizations, and for good reasons. I don't see the need for 'one size fits all'. That doesn't mean that we shouldn't try to make similar data sets more compatible, but often data was collected for different reasons and by using slightly, but crucially different experimental designs. Furthermore, research and knowledge is always changing and today's categorization might have to be broken up tomorrow.