Stuff like this always makes me wonder: why? They had to put in the effort to write the code to do this, to no clear benefit (it reduces apparent connection latency, I guess) and virtual certitude of incorrect behavior. It’s not laziness, because it’s additional work; it’s not run-of-the-mill incompetence because the implementation, while completely wrong-headed, seems to function “correctly” under the apparently desired behavior. The only thing I can think is that they’re attempting to game network performance benchmarks, which is outlandishly arrogant and incredibly stupid.
This is a very good writeup, by the way. It’s clear, concise, and has just the right amount of snark.
Stuff like this always makes me wonder: why? They had to put in the effort to write the code to do this, to no clear benefit (it reduces apparent connection latency, I guess) and virtual certitude of incorrect behavior. It’s not laziness, because it’s additional work; it’s not run-of-the-mill incompetence because the implementation, while completely wrong-headed, seems to function “correctly” under the apparently desired behavior. The only thing I can think is that they’re attempting to game network performance benchmarks, which is outlandishly arrogant and incredibly stupid.
This is a very good writeup, by the way. It’s clear, concise, and has just the right amount of snark.