In a talk on innovation at the Berlin-Brandenburgische Akademie der Wissenschaften Juergen Mittelstrass, one of Germanys great living philosophers, added one more slot to the well-known classification of science, leaving us with: 'fundamental research', 'application-oriented research' and 'product-driven application-oriented research'. While he was obviously trying to grant some innovative-potential to research formerly known as 'applied', it soon became clear that innovation would be found rather in the first than in the last box and the terrain for the unexpected was shrinking as it is obvious that innovation by nature is nothing you can plan for and the application-pull would lead to optimizations, inventions, solutions but not to surprises.
His critique of todays ever-growing emphasis on applicability in research-funding was massively amplified by Harald zur Hausen (Nobel-prize in medicine 2008), who reminded everybody that his ground-breaking results were never possible would he have been forced to do that agonizingly predictable research that is on every politicians agenda today. Under the guise of the responsibility to 'give something back to the taxpayer', funding schemes become more and more excel-sheet-driven. The demand for fine-grained 'milestones' with clear 'expectables' and costs over the whole funding-period (of typically three years) already at the time of grant-application results in scared 'chicken-science' without much space for the unexpected that is one hallmark of innovation. This 'accountability' (that is measured with static parameters such as number of publications, patents, invited talks,...) together with the need to get results that are common-sense in a tightly knit scientific community is killing creativity, innovation and ultimately the return on investment of taxpayers' money, eliminating the central argument for the bureaucratization of science by science-accountants.