I've been lucky enough to attend one of Peyton Jone's talks in person. He is not suggesting to "write a formal spec" before you do research. In contrast, his attempt is to be agile: keep going back and forth in the paper writing as you do research. Only then you will see what's missing.
Essentially, what he is suggesting is actually "agile academic research", compared to the "waterfall academic research" (do stuff, then write-up a paper with what you've got...)
Comments
I've been lucky enough to attend one of Peyton Jone's talks in person. He is not suggesting to "write a formal spec" before you do research. In contrast, his attempt is to be agile: keep going back and forth in the paper writing as you do research. Only then you will see what's missing.
Essentially, what he is suggesting is actually "agile academic research", compared to the "waterfall academic research" (do stuff, then write-up a paper with what you've got...)