Scientific Research: Moving beyond bubbles to platforms.
The Economist contrasts the current start-up culture to that of the dotcom bubble. Their special report argues that while the dotcom bubble was characterized by heavy investment in a single idea with a “Build it and they will come” attitude, the current generation of startups is characterized by tinkering and rapid experimentation. They argue that this is possible thanks to the cheaply available “platforms” of open source software, cloud computing, APIs, and social media dissemination. And they argue that this time, it is no bubble. That it will eat the world. That these lessons apply far beyond tech startups, to businesses in general and even governments. The special report is well worth reading in full.
So how about science?
After all, experimentation is what science is all about. Yet the platforms that run that world are largely absent from scientific research. Research looks more like the dotcom model, where each research team must make an immense investment in infrastructure and data gathering up front. Must hire expertise in-house for each step of the process, from experimental design to data collection to analysis to writing. Such large teams resemble more the companies of previous decades with their own development, marketing, and research departments, then the agile start-ups currently building the future. The resulting products are monolithic, “build it and they will come”.
I believe that this is the cornerstone of the open science and reproducible research movement.
Update (2014-07-24)
This seems to be exactly what Martin Fenner is talking about in “Roads, not Stagecoaches” – the value of transformative infrastructure over individual software contributions.