Why is the new syntax so clunky?

thing myThing = new thing;

When it could be:

new thing myThing;

ixid:

> Why is the new syntax so clunky?
>
> thing myThing = new thing;
>
> When it could be:
>
> new thing myThing;

Sometimes you want to write:

class Foo {}
class Bar : Foo {}
void main() {
   Foo b = new Bar;
}

Otherwise there is 'auto' too:

auto myThing = new thing;

Bye,
bearophile

> class Foo {}
> class Bar : Foo {}
> void main() {
>    Foo b = new Bar;
> }

Thank you, could you explain what is happening in your example? Bar is inheriting from Foo, what are you getting when you create a parent of type sub-class compared  to Bar b = new Bar; and Foo b = new Foo; ? Foo b = new Bar won't compile if you add members to Bar and access them.

On Monday, 7 May 2012 at 17:52:01 UTC, ixid wrote:
> Thank you, could you explain what is happening in your example? Bar is inheriting from Foo, what are you getting when you create a parent of type sub-class compared  to Bar b = new Bar; and Foo b = new Foo; ? Foo b = new Bar won't compile if you add members to Bar and access them.

Hello,

Foo is the "interface" you'll have to bar.

So, a bit of a bigger example:


-=-=-=-
import std.stdio;
class Foo {
    void doStuff() {}
}
class Bar : Foo {
    void doStuff() {
        writeln("Hi");
    }
    void doThings() { }
}

void main() {
    Foo f = new Bar;
    f.doStuff(); // prints "Hi" to the screen
    f.doThings(); // doesn't compile
}
-=-=-=-

So, as you can see, if you have a Foo, you can't call "doThings" using it. However, if your Foo is actually a Bar underneath, then it'll use Bar's version of "doStuff".

OOP isn't terribly hard, but I suggest reading up on it some to grasp the concepts (and especially so you can see the benefits).

Here's a link that might help you get started on some of the fundamentals:
http://en.wikipedia.org/wiki/Object-oriented_programming#Fundamental_features_and_concepts

On Monday, 7 May 2012 at 18:03:45 UTC, Chris Cain wrote:
> On Monday, 7 May 2012 at 17:52:01 UTC, ixid wrote:
>> Thank you, could you explain what is happening in your  example? Bar is inheriting from Foo, what are you getting when  you create a parent of type sub-class compared  to Bar b = new  Bar; and Foo b = new Foo; ? Foo b = new Bar won't compile if  you add members to Bar and access them.

> Foo is the "interface" you'll have to bar.
>
> OOP isn't terribly hard, but I suggest reading up on it some to  grasp the concepts (and especially so you can see the benefits).

 D also takes the approach with classes and objects that is closer to java than C++: All methods are overridable for polymorphism by default. Using interfaces and safely up&down casting using the known interfaces is so much easier. I haven't done much with C++, but these types of things are more error and bug prone than D and java; Not to mention the STL gives me a total headache.

 I have a good mental idea of how to explain OOP and inheritance using a RPG-like game example, which would be less confusing than the usual animals and shapes (At least it is to me).