On Wednesday, May 20, 2015 at 11:27:56 AM UTC+5:30, Steven D'Aprano wrote: > On Wednesday 20 May 2015 14:30, Rustom Mody wrote: > > > And what about Java? > > http://stackoverflow.com/questions/166033/value-semantics-and-pointer- > semantics > > Congratulations, you have found yet another example of the Java community's > collective delusion and insistence on misusing terms for the maximum > confusion possible. > > The mental gymnastics they go through to force the round peg of pass-by- > sharing object semantics into the false dichotomy "pass-by-value versus > pass-by-reference" is just crazy. > > One consequence of this is that the meaning of "call by value" depends on > whether the value you are talking about is a boxed or unboxed value. Unboxed > values are copied. Boxed values are not. Except that they are, if you > understand that "the value" doesn't refer to the actual value, but to some > hidden and implementation-dependent reference to the value. > > To quote Fredrik Lundh: > > well, I guess you can, in theory, value an artificial number > assigned to an object as much as the object itself. > > "Joe, I think our son might be lost in the woods" > "Don't worry, I have his social security number" > > > http://effbot.org/zone/call-by-object.htm > > > Rustom, if you are serious about this approach, then the implication is that > if I execute "x = 23" in Python, and I ask you what the value of x is, you > should answer something like "146588120" (that's the implementation > dependent "value", i.e. the address of the int 23). > > It's just *nuts*, and all because the Java folks are unwilling or unable to > distinguish between the language semantics and the implementation of the > language. And you're falling into the same hole.
I am not arguing for the java-folks terminology or outlook. I am arguing that for languages that execute on von Neumann machines memory is not an optional negotiable concept. And pointer is just first-classed memory as lambda is first-classed function. Some languages firstclass it -- most notably C, also C# which increasingly fascinates me in its choices. Others try to to unfirstclass , ie hide it in various half-assed ways -- Python, Java, Lisp. Others try to hide it in the most heavy-handed way possible -- haskell But you can never really hide it [law of leaky abstractions] eg in haskell one makes an infinite list of ones thus ones = 1 : ones Now if you generalize this to repeat x = x : repeat x and then do ones = repeat 1 you get poorer performance because the ones makes a box-n-arrow loop which the repeat does not achieve. To get round that we do repeat x = repeatx where repeatx = x : repeatx In short any language that is implemented on von Neumann hw will need to address memory. The language-designer may believe that there is a choice to 'unfirstclass' it. However people discussing language have no such choice: If you dont have it in the formal implemented language, you must have it in the informal discussion-language [With a large amount of wild waving of hands] <Admission> I dont like teaching this. viz that in python x = [[1,2],[1,2]] is equal to y defined as z = [1,2] y = [z,z] And although they are equal as in '==' they are not equal as in behavior, memory usage etc, a fact that can only be elucidated by box-n-arrow diagrams. So as far as I can I teach by tabooing extend,append and all such mutablers In the end though this is cheating </Admission> -- https://mail.python.org/mailman/listinfo/python-list
