[DGD] Clonables versus Inheritables

Josh Dady jpd at indecisive.com
Wed Aug 10 16:48:01 CEST 2005


On Aug 9, 2005, at 11:27 AM, Noah Gibbs wrote:

>> The kernel's rules prevent multiple compiled
>> versions of the same object from existing at the same time.
>>
>
>   This isn't entirely true.  DGD's rules (not specifically the  
> kernel's) keep
> multiple compiled versions of the same object from existing at the  
> same time,
> unless N-1 of those versions (called issues of the object) are  
> destructed.

I was specifically avoiding terminology that assumes he already  
understands the concepts (i.e., "issues" of objects).  If you load an  
object, make a clone of it, and destroy the master, the "compiled  
version" (i.e., the bytecode) still exists (otherwise the clone would  
stop working).  If you edit the source file, load the object again,  
and clone it again, two different "compiled versions" exist.  That's  
what I was referring to (the same would be true without editing the  
source file, but that's beside the point).

>> Any
>> inherited object can't be upgraded, so the master must be destroyed
>> and re-created.
>>
>
>   You mean the *parent* must be destroyed and re-created.  Sorry to  
> be picky,
> but "master" actually means the object from which a *clone* is  
> created, so this
> is an important distinction in this case.

That probably should have been "...can't be upgraded; it must be  
destroyed and re-created instead", although the distinction would be  
lost on anyone who hasn't read the glossary (see above).  Even then,  
an object's "parent" could mean anything (I generally only use  
"parent" when I'm talking about a directed graph of some sort).

--
Joshua P. Dady
http://www.indecisive.com/




More information about the DGD mailing list