We live with the bold categorization of research as being either 'fundamental' or 'applied'. The emphasis in funding - and broadly in the public understanding - is on the supposedly more valuable *applied research*.
Scientists engaged in fundamental research, on the other hand, are widely seen as geeks, as nerds in ivory-towers of academia, kind of wasting taxpayers' money for their personal entertainment, dabbling with expensive machines, finding ultrafast neutrinos and dismissing them again...
At the same time innovation is imperative. So innovate we do. All the time.
But seriously, what kind of innovation could we expect when we are asked to do research on optimizing the rubber of a tire, or if coerced to develop a better mp3? What can we expect if somebody pays our research to
make cars more fuel-efficient? Certainly there would be some neat progress. Some nifty inventions. But innovation?
Let's look back. What would we have gotten when, 30 years ago, we would have researched how to make light-bulbs smaller and less energy-hungry? We might have gotten smaller light-bulbs, with some trickery inside. Would we have light-emitting diodes today? Probably not.
Anybody remember those tubes in old receivers? Doing applied research on them would have led to what? To the invention of a transistor?
Would we have discovered and understood x-rays by looking for methods to study bones in a living organism early 1900? Or take electromagnetic waves. Are you sure we would have encountered them during our quest to invent some type of long-distance communication 100 years ago? I bet carrier pigeons would have made the race.
To put it the other way around: what fundamental discovery concerning laws of nature did *not* result in a multiple gazillion dollar market? (yes there are a few - but you get the point). The danger of doing research in a system that demands predictability of results, as well as clear, controllable project-plans with well-defined milestones and a written concept for IP- and technology-transfer at the end is that you mostly get exactly that: predictable results.
Don't get me wrong, those results can be great, helpful, important. But the request to do only the predictable will yield the predictable - at best. Stubborn application-driven research yields incremental inventions and a predictable but very limited return on investment.
The real disruptive stuff can not be planned - by definition. If you aim for real invention it is necessary to look for some deep understanding of nature - and then get somebody inspired to create an application from that. As also industry is demanding to be 'less nice' if you wish to be innovative, science should withstand the pressure to serve as an incremental problem-solving machine.
The return on investment will be much bigger.