Chesterton's fence has long been a staple of HN discussions. Unfortunately, it involves no risk/reward analysis and it invariably leads to needless conservatism. That's probably not a big deal in medical research where risks are often high, but it's a problem when applied as a general principle.
(The opposite wisdom is probably grandma's ham [1]. Consider that if grandma had passed away, they might never be allowed to try cooking the whole ham.)
The point of Chesterton's fence is cost/benefit analysis. It's not saying the fence mustn't go; it's saying you need to know why the fence was put there in order to evaluate whether it's best removed or left in place - that is, your analysis needs to be informed by understanding of the status quo in order to be accurate.
Grandma's ham is a great example. It might have been some subtle trichinosis-related failure case that cutting the end off a ham reliably prevents, and it's a shame we had to lose your great-uncle Joe to find out about that. Or it might have been an issue of pan length. You can't know until you ask, so you ask if you can. If you can't ask, then sure, you do the best you can with what you have - I should like to hope there are no blind dogmatists here. But if you can ask, you'd be a fool not to.
(The opposite wisdom is probably grandma's ham [1]. Consider that if grandma had passed away, they might never be allowed to try cooking the whole ham.)
I quite enjoyed that blog post, btw.
[1]: http://www.angelfire.com/ma/artemis9/humor/joke7.html