Breakthrough! How many times have you seen a science story with that word? A university professor, sometimes a startup company, announces that they have found the answer to a problem getting in the way of better batteries, production of fuels from sunlight and spiderweb, and don’t get me started on health news!
These stories are interesting for a wonk like me. Some of them, particularly the ones on sunlight and spiderweb, impinge on research I’ve done, and it’s fun to see where they fit in that chemical universe. But they are not breakthroughs and do little more to advance public knowledge or utility than the clicks they provide their website.
University research, and many great ideas that startups love, are nowhere near practical use. What they are touting are results obtained on a few samples under very constrained conditions in a laboratory. Things probably didn’t work out consistently for a long time. More responsible researchers will wait for that consistency before announcing. Many don’t even do that.
The announcements usually aren’t erroneous as to the science, although they can get hashed up through combinations of professors and university public relations departments trying to popularize difficult concepts, filtered through reporters who don’t know the questions to ask. What they miss is that there are many time-consuming steps from lab bench to everyday use, and many technologies fail along that path. And that’s just turning science into engineering into product. If it’s something that needs regulatory approval, like a drug, it will take longer.
Here’s an example of the kind of thing you would see in a technology that is getting closer to use: Big equipment, workers in protective clothing, as in the photo above. And there may be yet another step up to full production.
A related problem is that of software. Its lifecycle isn’t quite the same as from lab-scale chemistry to a product, but maybe it should be. As software becomes part of more essential systems, it has the potential to do great harm. So maybe more of a product development cycle, with more quality control, is in order. This article ponders whether software developers can really be called engineers. I think they could if they were subject to the constraints that engineers are.
Update (November 6, 2015): Here’s a good example of how hard it is to bring something from an idea to a useful product. This article lists all the things that can go wrong. And there are more when you’re starting from a laboratory breakthrough.