Attaching a gift of money to a birthday card, for one
Yes, and then the birth card and the money/check is gone as some post man just keeps it for himself. Same with cash.

No one in Europe is going to send checks/cash via mail, unless they are special checks that can only be cashed in by "the owner/the addressee" showing a passport.

Even if a bank offers a wire transfer for no fee, you still have to know your recipient's bank account number, and you still have to either "go to the bank IN PERSON" or subscribe to cellular Internet service to set one up.
No, you don't have to subscribe to a cellular internet service.
That is a pretty dumb assumption. Any internet service does it. And ofc you can mail the transfer order to the bank, so you have no need to go there in person, and ofc. that service is free for individuals.

Bottom line: payment habits are a cultural thing.

Some have a donation box near each of the auditorium's exits; this is the common practice for Jehovah's Witnesses. Others pass around a bag or tray into which members of the congregation drop cash or checks
Yes, I know. The point was: is actually someone putting money into that? I know no one who ever did, but anyway, I'm an Atheist, I'm only in the church if some get married or a child is baptized.

The parent was wrong.

Ofc there are still checks in use, in some countries more than others. (E.g. France and UK)

how do individuals pay other individuals through the post
You don't. Why would anyone do that? You use a wire transfer. The only checks I ever get via mail are from my old landlord when he refunded excess utility costs (water/garbage): AND I HATE IT, as I have to go to the bank IN PERSON to draw them. When I don't draw a check fast enough he even calls me and insists that I do so soon :-/

And how do churches collect donations?
Cash or wire transfer. (And actually, are there really people giving donations to churches? Shudder!)

Everything that can be treated by "stimulating" the vagus nerve indicates that there is a problem with the vagus nerve. And that is in 99% of all cases a blockade/pressure by a muscle or sinew pressing on the nerve somewhere along the path that nerve is running.

And that can be treated easily with:
o heat
o massage
o simple herbal medicals like Camphor/Arnika
o Tai Chi
o Chi Gong
o Osteopathy
o Chiropractics
o Shiatsu/Accu pressure massage

Depending on where the blockade is and how difficult to treat it is.

Main reasons are: very bad body posture and lack of sports and physical activity.

I doubt an implant can long term fight against those two main reasons for illness/malfunction.

Heck, even a simple dance class where people learn to relax the hips and stay erected with relaxed shoulders would likely solve more than 50% of the cases suffering from vagus nerve blockades.

You neither defined any of the four possible constructors, nor the destructor... and this means you get the defaults. And this has nothing to do with 'when needed', only with 'did I provide one myself, so the others are compiler generated or not?

You are mistaken. If I don't define a ctor, the class has no ctor, hence its members are never initialized. They get the values the memory has at the point where they get allocated. For statically allocated variables that might be a zero initialized area. That means if I new B() somewhere it gets what ever data is on the heap or if I do B b; in a function, the object holds what ever random data is on the stack.
Because: it has no ctor!!

However as D is inheriting from B and has a ctor the compiler is auto generating the "missing ctor" of B, because D::D() {} wants to call that ctor B::B(). That auto generated ctor for B is usually only accessed where ever D is used. But I already had cases where (IMHO a bug) where the construction process of B suddenly changed (I mean declaring a variable like this: B b;) just because D was "known" or instantiated.

In other words: objects of type B are instantiated differently depending on the fact if they are a part of a D object or if they are solitair.

I suggest you read the C++ definition about when ctors are "generated when needed".

Of course I fully agree with you that variables should be initialized :D But it is easy to forget a field in a ctor, no idea if modern compilers give warnings about that (and to lazy to try that now on my Mac).

Do you really want to tell me that no other language has less used features, or pitfalls, which are simply irrelevant when good coding styles are used?
I did not say that. But people claimed C++ would be an easy to learn language, and I like to point out: the contrary is true.

Your way of posting implies you don't know what an API is.

E.g. calling a middle ware an API is nonsense. And there is certainly no "Java API" for that.

To implement middle ware we in our times use: SOAP, REST, various message queues, CORBA and RMI, and probably custom made ones.

So: YES, if you want to call that an API (the correct term is: "a standard") then there is for most of them only one API, in other words: the CORBA binding for SUNs/Oracles Java and IBMs Java and HPs Java: are identical! That is the point of having such an API.

Are you also implying that there is no creativity and design involved in designing an API?

How do you come to that conclusion? I did not say that. Nor did I imply that.

When I ask 12 (competent) developers as you proclaimed to define an API for reading/sending/processing eMails like this: https://javamail.java.net/nona...
then I expect them to get at max three variations of such an API which should look very similar.

After all there is no sane reason to craft an API that diverges much from suns original work.

You might disagree, but you would be wrong.

Same for any other API. If you find it complicated to design APIs then perhaps restrain yourself from it.

Just because there are multiple ways to implement an API ...
Implementation is not defining an API!!!
... does not mean there is only one API for a given area of computing. Of course not, but there is one mainstream and a few rarely used ones, like SAX versus Schema based or RelaxNG based XML parsing/processing. In Java you have on API and multiple implementations.
Same for JDBC, JDBC is the API and various database vendors implement that API, but there is no "other version of JDBC than original JDBC because it is for funk sake not needed!!

If you asked a dozen programmers to independently design a dozen APIs for the same area of computing, the result will be a dozen different APIs.
Then minimum 11 of that dozen programmers are incompetent.

Seems you are mixing up API with implementation anyway in your post.

APIs (as well as DB schemas) are not copyrightable, for a reason.

APIs can't be copyrighted.

Only the implementation of said APIs can.

The text of the article is wrong in that regard or simplified.

All points you made are wrong :D

So, you had failed a simple job interview about C++ based on wrong answers regarding constructors, wow ... and you still think the language is easy, when you are not even aware that the member variables are not initialized at all when there is no constructor?

I did not define a virtual destructor as it is irrelevant for this example.

Good, now you know the member variables of B are not initialized at all, dare to try a new answer ... considering that I already pointed out "there might be a difference depending on compilation units" ???

My bet: your answer will be wrong again ... but no problem. C++ is such an easy language you pick this particular problem up quite quickly :D

Do you remember a language name? I'm looking for an interpreted MI capable language to port to iOS and use as scripting language for an App.
Any idea?

Perhaps read a book about the Taligent Frameworks?

When do you use (single) inheritance? When you "extend" a class or when you "specialize" a class? Whenever you want to extend a class in C++ (lets name it B) you should not inherit from it but craft a new class (lets name it N), worst case via a mixin template construct, and compose the desired class by multiple inheritance so that D is specializing B and N.

That way you can reuse N without being dependent or messing up D or B.

The simplified single inheritance and interface model caters for all practical examples, and *vastly* simplifies the internal workings. (X)

No, it does not. As you have to program the implementation of everything you "inherit interface wise" multiple times everywhere where you "implement" that interface.

If you want to use simple delegation (via IDE code generation etc.) the delegatee loses the original this pointer ... (X) above No simplification but a complication.

How large is a pointer to a member variable?
That is a stupid question. Yes, I know there are no stupid questions, only stupid answers. However asking questions that don't belong to the context: is stupid.
Answer to your question: implementation dependent. What exactly do you mean with a pointer to a member anyway? A true C++ member pointer? Like &classname::member? Usually as big as any pointer on the given platform. But can be optimized as small as a byte, however I'm not aware of any compiler that does that.
If you mean &object.member, a safe bet is: the size equals the typical pointer size on the given platform.

I did not argue about "shared_ptr", that was someone else.

The article you linked is bogus. I stopped reading after this: Ownership of resources is a central part of software engineering.

That is plain wrong. The idea of "ownership" was introduced with C++ ... it is a completely flawed idea to approach resource handling especially memory management.

Example:
class Person {
Person *spouse;
}

Who owns whom, when to delete the spouse? Obviously the idea that one person owns the other one makes no sense in terms of C++ ...

The answer is, if you use an uninitialized variable, the results are undefined.

Your answer is wrong.

I suggest to read a book about C++ and the mess about constructors.

If you ask politely I explain you what is going to happen depending on where you use either B or D ;D

So you never tried the -auto-fix option?
Shame on you! ;-)

C++ is harder than other languages.

E.g. "a constructor for a class is auto generated by the compiler when needed"

Under what circumstances does this rule apply?

Oh, and what are the pitfalls?

E.g.

class B {
int i;
float f;
}

class D : public B {
int s;
public:
      D() {
          s = 5;
      }
}

B b;
D d;

What value has b.i?
What value has d.i?

Hint: consider B and D to be in different compilation units.

My bet: your answer is wrong.

Wow: that above was a super simple example about problems in C++ that make C++ a very hard language

Actually it goes like this:

C: You shoot yourself into the foot.
C++: You shoot yourself into the foot, and have blown of your whole leg.