R help - Nov 2009 - Question about S4

Dear R-ers,

I don't understand the following, maybe someone will help me 
explain:
> setClasss('A')
[1] "A"
> new('a')
Error in new("a") : 
  trying to generate an object from a virtual class ("a")
> setClass('b', contains='a')
[1] "b"
> new('b')
An object of class ?b?
<S4 Type Object>

In what way is B more concrete than A so that it's possible do 
instantiate B but not A? I don't quite get it. B adds nothing to 
nothing, and yet it's instantiable, while it's base is not. Makes 
no sense to me.

-- Hun

PS. All class names were upper-case, I messed up while copying the 
code, but it has no effect on the result. Thanks for help.

--Hun


On Mon, 23 Nov 2009 10:28:37 +0100 "Hun S. Tesatte" 
<hunsyntesat at hush.com> wrote:
>Dear R-ers,
>
>I don't understand the following, maybe someone will help me 
>explain:
>
>> setClasss('A')
>[1] "A"
>> new('a')
>Error in new("a") : 
>  trying to generate an object from a virtual class ("a")
>
>> setClass('b', contains='a')
>[1] "b"
>> new('b')
>An object of class ?b?
><S4 Type Object>
>
>In what way is B more concrete than A so that it's possible do 
>instantiate B but not A? I don't quite get it. B adds nothing to 
>nothing, and yet it's instantiable, while it's base is not. Makes 
>no sense to me.
>
>-- Hun

I'm curious about why no one has answered my question below. I 
can't imagine it would be because no one knows how to answer, it 
must be something basic I am ignorant about. But I have never seen 
such a pattern, it seems strange to me that a class with an empty 
definition is automatically virtual, but a class extending it 
without adding anything is not. I am really puzzled, there must be 
some design decision behind this, but I can't figure out it's 
purpose and usefulness.

I'd really appreciate an explanation. Thank you.

-- Hun


Dear R-ers,

I don't understand the following, maybe someone will help me 
explain:
> > setClasss('A')
[1] "A"
> > new('a')
Error in new("a") : 
  trying to generate an object from a virtual class ("a")
> > setClass('b', contains='a')
[1] "b"
> > new('b')
An object of class ?b?
<S4 Type Object>

In what way is B more concrete than A so that it's possible do 
instantiate B but not A? I don't quite get it. B adds nothing to 
nothing, and yet it's instantiable, while it's base is not. Makes 
no sense to me.

-- Hun