Frank,

I think we do not disagree at all. I always deliberately write 'object reference variable', thus making clear that an O-type variable is always merely a reference to the object, not the object itself.

I already explained to Dragan that he kind of mixed up me and Naomi when he started to talk about object referencing. When you go back in the thread, you will see that I pointed out to Naomi that it is, normally speaking, better not to change the value of a passed variable. This is also true for variables that refer to an object.

>Pater,
>
>>As I read it (and I did run the code), the first run does not remove the form, indicating that it is a pass by VALUE
>
>The memvar oForm was passed by value (copy of), but it's a reference (address of) to an object otherwise we could not change the properties of the underlying object. When that code set the memvar to null it was only affecting the local stack copy of the memvar oForm, not the real oForm memvar.
>
>>while the second run is indeed a pass by reference.
>
>The address of the memvar oForm was passed to the routine, it still references (holds the address of) the real form object so we can change the form's caption. But now when the loObject was set to .null. we are actually changing the oForm memvar, which decrements the reference counter of the form object to 0 which makes the form destruct.
>
>>So, apparently objects (no, better: object references) are not always passed by reference, but can also be passed by value. Is this confusing?
>
>The way you state it above is not correct.
>
>You always have an object by reference. We really don't have a native way to pass object copies around in VFP like can be done in other languages.
>
>The memvars that hold an object reference may be passed by value or by reference.
>
>>>Peter, et al
>>>
>>>Objects are always passed by reference. The memvars pointing to them may or may not be passed by reference. A subtle difference: