I think it's more of a guiding principle rather than a steadfast rule. Like I would think "releasing code to production" is a one-way door. Even though it can b

I think it's more of a guiding principle rather than a steadfast rule. Like I would think "releasing code to production" is a one-way door. Even though it can be undone, it can have pretty bad consequences and breaks existing protocols. On the other hand "researching a new topic" cannot be technically undone, but it is a two-way door because there is not any harm if it goes wrong and you're working in the best interest of your team.