Unfortunately, that's not entirely accurate. The problem with language level benchmarking is that the benchmarks are only as good as their implementations and while I agree that different languages encourage different kinds of idioms it seems only fair to use the same data structure/technique when the language offers that choice (as Rust does with intrusive data structures).

The Rust program was not a faithful reimplementation as the author concedes. I expected a more closer comparison given the somewhat broad title of the article. Claiming the perf benefit came out of the thinking encouraged by the language is okay but debatable.