Most of the .NET literature I've come across reads somewhat easily because it is so similar to Java and I had already taken several Java classes.

But one of the things that I find confusing in the .NET literature is their talk of boxing and unboxing and their emphasis on Value Types and Reference Types.  I couldn't remember a similar topic in Java.  What I do remember from Java was that almost everything is an object except for certain native (intrinsic or primitive) types like short, int, float, etc.  For efficiency, these were not objects but there exists wrapper classes for them (like Integer etc) that can turn them into true objects.

Is this basically what .NET's boxing and unboxing is?

.NET literature says that everything is an object.  But it's sort of confusing.  Value Types get created on the stack and Reference Types get created on the CLR heap and boxing basically makes a copy of a Value Type into a Reference (CLR heap) Type and vice versa.  Most of the time, boxing and unboxing is done automatically in the background.

To me, Value Types sound a lot like Java's intrinsic type and boxing is the same thing as providing a wrapper class to turn it into an object (Reference Type).  This is because in Java an intrinsic type is created on the stack and an object is created from the Java heap.

Is this automatic boxing and unboxing what allows Microsoft to say that everything in .NET is an object?  Since it automatically wraps Value (stack based) Types into Reference Types (CLR heap), the programmer can think of everything in .NET as an object.  However, there are situations where the programmer has to do boxing or unboxing manually.

I sometimes wonder if Microsoft was playing a semantic game so they could make marketing claims that everything in .NET is an object (and hence give it a marketing edge over Java).  I personally find Java's concept of wrapper classes an easier concept to understand and wish that Microsoft had done the same thing in .NET.

Or am I confused about this topic?