Re: Dynamic languages (again)

Quoting Jim Haungs <[email protected]>:

> There's nothing inherently reflective about dynamic languages, nor vice
> versa. The concepts are orthogonal.
Yes, both concepts are orthogonal in theory. But is there some reason in practice for developing a dynamic language other than supporting advanced reflection capabilities? The only exception I can think of is to create a language meant for interpretation only (for instance, the scripting languages used for Web programming).
 
> Dynamic (at least in my definition) means that variables are bound to values
> at run-time, and methods are dispatched via run-time look-up. True, this
> requires the VM to introspect the method dictionaries, but nothing requires
> the running user-level program to actively introspect anything.
I completely agree. Indeed, for most scenarios, you do not need to use reflection at all. But the possibility of intercessive reflection is an advantage of dynamic languages, in my opinion.
 
> And Java has reflection without fully-dynamic typing. By "not fully-dynamic"
> I mean you can declare a variable to be of type Object, and put any reference
> type in there, but you can't put an unboxed primitive in there, something
> you can do in Smalltalk and Ruby because everything inherits from Object,
> even so-called "primitives."
Yes, Java has reflection, but only introspective reflection. Intercessive reflection makes it unconceivable (for me, at least) the use of a strongly typed language. Or is there some way to have intercessive reflection with a "static" language? (It would be nice to learn how it could be done, and which mechanisms for intercession are possible in such languages.)

I am not saying that dynamic languages are bad. Indeed, I really enjoy Smalltalk and the ideas behind it (such as the use of reflection to change the development environment), and I have even done some programming in it. Smalltalk is very cute, and Eiffel (a static language) attracts me a lot too. What I mean is that the choice of the language designer is subject to trade-offs: you gain in one side and lose in another side. You can choose some extra mechanisms for getting a little more reliability or you can choose some extra features for getting a little more flexibility. Anyway, test-driven design is strongly recommended. So, any affirmation that a dynamic language is irrevocably better than a static language (or vice versa) for the same paradigm is not to be seriously considered.

Antonielly Garcia Rodrigues

__________________________________________________
AOSD Discuss mailing list    -    [email protected]
To unsubscribe go to http://aosd.net

Check out the AOSD.net Wiki: http://aosd.net/wiki