You could always just return a tuple, then use tie on the caller side. To use your example,

std::tuple SinCos(float angle) {
...
return std::make_tuple(sin, cos);
}

float sin, cos;
std::tie(sin, cos) = SinCos(angle);

Also, your last question isn't really a question, is it?

The assembler? Then how do you intend to compile anything into machine code? Must be quite a project you have there

Yes, that's what the PIMPL idiom is for. Here's a nice introduction to it .

As with most everything, there's a way to solve that problem in C++, it just takes some work and knowledge to know what, when, and how to use it. I do agree that the language is too big though; conceptually, C++ could be broken into 4 distinct components (C, STL, templates, OOP), as each have their own quirks and idioms that don't necessarily carry over well to the other, and some are even Turing complete on their own (ie template metaprogramming)

I'm not quite sure I understand your gripe properly. You're saying that since C++ has classes, and classes are defined in header files, that calling code is dependent on the class because... it's in a header file?

If it bothers you so much, use the PIMPL idiom. Or forward declarations where you can. Or template your calling code and write to static interfaces. You've got plenty of options, you just need to know how to use them. C++ is admittedly a very large (arguably too large) language, which is both it's most compelling feature and it's biggest drawback. You can use it for pretty much everything, but fully grasping the language and knowing when to use what feature takes time. This is partly due to trying to maintain backward compatibility with C, while still incorporating new language features. Because of this hodgepodge, iterative development, some dubious choices have been made. D is meant to take the good parts of C++ and cut out the bad; it is a good idea, and a lot of the foremost C++ experts are D supporters. It's a shame that D will probably never take off.

If you're going to complain about C++, at least use valid complaints. Like about how 2 primary aspects of the language (templates and OOP/dynamic polymorphism) are partly incompatible, or about how there's no dynamic multi-dispatch built in to the language still, or how there's no concepts, or how C++11 async doesn't really give task parallelism. (And yes, there are others, many others). But classes and header files? Cmon.

And yes, the famous, fictional interview...

I was citing your post. You're the one who used them as an exemplar. Why are you asking me this?

I have no idea what this question is referencing

I provided no judgment or commentary about whether I agreed or not with the interview procedure or questions. My point was merely that being able to operate on the bit-level is still useful, mandatory in fact, to work at the places you cite as not needing them. YMMV

It's funny that you mention that. I've recently had to perform some interviews (as an interviewer), so I was looking around the 'net for some tips on how to be a good interviewer. Lo and behold I came across this piece by Steve Yegge, about the kind of questions he asks as an Amazon interviewer.

He lists "The Five Essential Questions" for phone interviews , one of which is, as you would say, 'bit bashing'. The full list is:

So while you might not use bit bashing after you start working for those companies, you'll never get the chance to if you don't know how.

Adds hiding? You mean... encapsulation?

Is that the only thing you can think of that C++ adds to C? If so, that may explain why you have such a poor view of C++

"If Java had true garbage collection, most programs would delete themselves upon execution"

– Robert Sewell

Ebola is only contagious when the symptoms are present , NOT during the incubation period. The symptoms of Ebola are pretty pronounced, so if you see someone projectile vomiting and bleeding from their eyeballs, steer clear, but otherwise you should be alright. From the WHO Ebola FAQ:

I was referring to Neymar specifically -- he got a fractured vertebrate in the game against Columbia and had to sit out the rest of the World Cup.

Of course, when you get kneed in the back and get a fractured vertebrae, not much acting is required.

4.8 was also the first GCC version that (more or less) had full support for C++11. It's not terribly surprising to get more regressions when adding (many) new features.

I read the article/rant, and he's bitching about the complexity and time required to learn and use modern frameworks and tools, not so much about planning and management. The summary is a bit misleading.

I think part of the problem is that "programming" is itself so diverse. If you were to be a graphics programmer, you would certainly need your linear algebra, geometry, etc. If you worked with scientific computing, you'd need even more math (e.g. differential equations, statistics, etc). If you worked as a DSP programmer, you'd need to know calculus (and then some). In contrast, web development doesn't really require any of these. However, they all involve "programming", and the people writing the software can all be called "programmers", even if one's writing a website (no math) and another is doing a fluid dynamics simulation (lots of math).

These vials are from over 50 years ago. Your comment would have made more sense if meant facetiously. From TFA:

Tell me again what role Obama had in this?