Word can refer to any context at any point in run-time.
Yes, but only 1 at time, right?
@secretis still a variable (taken from: http://ruby-doc.org/core-2.4.0/Binding.html ):
class Demo def initialize(n) @secret = n end def get_binding return binding() end end k1 = Demo.new(99) b1 = k1.get_binding k2 = Demo.new(-3) b2 = k2.get_binding eval("@secret", b1) #=> 99 eval("@secret", b2) #=> -3 eval("@secret") #=> nil
any-word!), and context is provided by separate values (
function!). Binding is a relation between two separate entities, which you can manipulate at run-time (words and functions/objects). Variable, on the other hand, is a way to reference a value, but not a relation between the two values.
@nedzadarek since "variable is a way to reference a value", "variable" (in your interpretation) can be either:
Word is not a variable in any of these cases, it never refers to the value directly, it's always refers to a context provided by some value (either by object on the heap or by function on the call stack).
did rebol have a bigger problem leaking context?
Currently Rebol 2 and Red are little different in this regard.
useseems like a solution in search of a problem in red
An interesting way of looking at it. I've long found it a very tidy way of creating micro-contexts. With Rebol 3's tighter control on context, I've found it a useful way of tracking word use—even within modules.
P: context [ F: [ P1: [ name: "John" ] P2: [ name: "Mary" ] ] ] set [list1 list2] compose [ (words-of (Context (P/F))) (values-of (Context (P/F))) ] ; this works set [list1] compose [ (words-of (Context (P/F))) ] ; this doesn't work forall list1 [ probe list1/1 ] forall list2 [ probe list2/1 ]
P: context [ F: [ P1: [name: "John"] P2: [name: "Mary"] ] ] set [list1 list2] reduce [ words-of context p/f values-of context p/f ] forall list1 [probe list1/1] forall list2 [probe list2/1]
composeapplication and you'll probably figure out that it's not what you expected.
compose/onlyinstead, which composes nested blocks "as-is" (though
composeis superfluous here, as
reducecan do the same job just fine). Moreover, I don't get why you need multiple parenthesis nestings.
One of the great and terrible things about Redbol langs is their chameleon ability. @nedzadarek, a great many details are hidden under a facade that makes Red quickly accessible to people, but there is the risk that they will see it "as just like X", when it really isn't.
That has been just my problem. The original REBOL core manual did not a good job on letting me forget about variables, pointer and classic OOP.