Yeah but I didn’t see anything about perf there.

I think it’s a valid point that the reasons why C++ is good for perf - compiled, statically typed, low-level, manual memory management, high degree of control over memory layout - would also apply to Rust. If they don’t then that would really call into question positive claims being made about Rust.

I was curious mainly because I remember some long comments he made about the relative value of linting and static analysis tools, specifically going into Coverity's analysis of the Doom 3 code (I think), fixing everything it flagged for some subset of the code, and then asking himself whether it was really better or more of a hindrance that obfuscated code.

IIRC, his conclusion was mixed: a lot of it was obviously beneficial and worth having turned on, but much of it wasn't, and going forward he intended to make it a limited but continual part of his toolchain.

So my interest in whether he'd tried Rust was whether he'd compared Rust's changes like the borrow checker against his earlier conclusions on writing good C/C++. Cute that he's tried and liked it, but I'd really like to see a more in-depth comparison from him.