Transaction rollback is probably what he meant. If that is a familiar topic to you the relation to undo should make sense and his point should be clear. If that doesn't make sense, look up ACID, Transactions, and Migrations in the context of a database.
There is a form of them which is invariant. You just have to write them correctly. Mathematics isn't as amazing as it sounds, there is a some ambiguity in how symbols are defined. Is your gradient with respect to rectilinear coordinates, or more general? Etc.
It's a Friday, quality tends to degrade during the weekend. Usually professionals reading the site tend to view and comment more during the work week while things are compiling/deploying. The weekend tends to be more hobbyist/lonely people like me.
The phrase occured only in the title and article and not in the comment they were responding too. It is possible they did not fully read the article(Most skim or read the title or title and abstract). In which case your response wasn't super helpful.
Not that I have studied this particular behavior, but evolutionary type theories like this one, tend to be found wanting after later research. For a more common example, look at the paleo diet. At a glance, the theory seems a bit bonkers if you consider natural feelings around kissing you may have had around puberty. I would imagine there might be a better explanation than some nuturing/feeding urge.
Edit: Not to be that dude, but for the downvote, explaining better how/why this feeding explanation makes sense would be nice. I thought my comment advanced the conversation somewhat.
The down votes were because you clearly didn't understand what preemption(or that this is really about it being a realtime flag) is or add much to the conversation(mentioning anything at all that might add something, like even a bad question, "I heard people using raspberry pis for embedded, would this help?"). The downvotes on the other guy are because asking for feedback on downvotes is discouraged.
To contribute, the PREEMPT_RT has been around as a set of patches(if not this specific one, then at least some version of making linux a realtime os.). If you need to know that something will run within a specified amount of time realtime matters.
EDIT: Don't use this and think it will help performance, generally it kills throughput.
Passing a "hardcore" coding test is something you can train for, and tends to not correlate well with actual software architecting and optimization knowledge.
You don't even know what 'hardcore' means in this context. Your literally judging their process essentially nothing and you have strong opinion on it. That said more about your prior then anything else .
I had the opposite experience when I met Jan Hesthaven at a GR conference. Turns out it is a German man's name and not a woman, which I found out after professing how much I love the hard 'J' versions book on spectral methods for time dependent problems.
