[racket] Pass by value/name/reference

From: Matthias Felleisen (matthias at ccs.neu.edu)
Date: Sun Jul 20 18:27:20 EDT 2014

call-by is a notion of time, pass-by refers to entity that is used (location if you so will). You can create a 2-dimension table of these options and you'll be able to find some language for almost any cell in this table. 

The common meaning of call-by-value is call-by-value/pass-by-reference. Both Java and Racket and a whole lot of languages use this combination. Call-by-name is also a common reference to call-by-name/pass-by-reference, see Algol 60 for an example. 

You can also create a calculus for any cell in this table. Google for Crank and Felleisen and POPL and you'll find something. With that I mean 

(1) a calculus whose basic axiom(s) capture/s the mechanics of parameter passing 
(2) and that satisfies Church-Rosser and has a Standardization Theorem (which defines a deterministic strategy for evaluation terms)
(3) and whose ST coincides with the conventional evaluator/abstract machine 

This holds for call-by-name, call-by-value (common meanings), in purely functional or imperative settings, and it extends to call-by-copy and several other mechanisms. 

The phrase "call-by-value is a reduction strategy" has no meaning per se but is a left over from the time when people hadn't figured out the above (pre 1070). 

It sounds like you're taking a course on PL and got some exercises to solve. I am glad we can help, and I assume you will give credit to the mailing list. 



On Jul 20, 2014, at 2:44 PM, Jon Zeppieri wrote:

> On Sun, Jul 20, 2014 at 5:02 AM, קוראל אלימלך <coral2301 at gmail.com> wrote:
>> Hey :)
>> How the arguments are passed in racket?
>> by value/ reference/name?
>> 
>> If u can add an example it will b good :)
>> Thank U !
>> 
> 
> I suspect two different things are being conflated here:
> 
> 1. the reduction semantics of the language
> 2. the question of whether or not data is copied at procedure-call boundaries.
> 
> 
> For (1), Racket's reduction strategy is call-by-value. As Jos Koot has
> demonstrated, in a procedure-call expression:
> 
> (fn-expr argn-expr...)
> 
> all of the sub-expressions will be reduced (that is, evaluated) to
> values before the function is applied.
> 
> For (2), Racket does not copy data at procedure-call boundaries. If
> you pass a mutable struct, the callee is able to make changes to it
> that are visible to the caller. This is unlike in C, where you would
> need to pass a pointer to a struct in order to allow the callee to
> operate on the same instance as the caller.
> 
> Unfortunately, "call-by-X" terminology is used to refer to both things
> and usually in ways that make matters far more confusing than they
> need be. See, e.g.,
> [http://en.wikipedia.org/wiki/Evaluation_strategy], which is just
> awful.
> 
> -Jon
> 
> ____________________
> Racket Users list:
> http://lists.racket-lang.org/users



Posted on the users mailing list.