[peterevans]

I think I understand Linus' concerns. I guess I'm not sure how you conclude that I do not, but, so it goes.

Languages do change, and implementations evolve. I think it's fair to temper criticism with the caveat that, hey, this was from a while back. And Linus' comments certainly were from a while back, considering the latest one was made in 2007. You didn't have shared_ptr, you didn't have unique_ptr, there was no auto type.

I'm not going to argue that these things make C++ awesome. I really don't work in C++ on a day-to-day basis. But the language is certainly different now than it was in 2007, and I think that does matter. At least a little bit :)

[drvdevd]

I think the number one legitimate concern I can see, after just reading this for the nth time, is the trustworthiness of C++ compilers, which I believe was not so great in the early 90s. Now, I have no evidence to back this up (yet), but I'm willing to bet that clang and LLVM are verging on as trustworthy as (if not more than) GCC at this point. Nevermind MSVC, which would require its own analysis. I'm not a C++ programmer, but you're right: clearly it has advanced. And Linus has gone off on GCC as well for producing crappy output as well (https://news.ycombinator.com/item?id=8089321)