Re: OOP vs imperative, was Hello v1.0.3 distributed programming language available (alpha)

"Nils M Holm" <nmh@t3x.org> writes:
>Anton Ertl <anton@mips.complang.tuwien.ac.at> wrote:
>> Smalltalk is an imperative language; e.g., it contains assignment and
>> its objects contain changeable state. What makes you think that
>> Smalltalk is functional?
>
>While I would agree that Smalltalk is not intended to be a functional
>language, I think the argument of assignment and mutable state is not
>valid.


Why not? It's how you know whether a language is imperative or
declarative.


>In "The Definition of Standard ML", Milner, et al describe ML as a
>functional language "in the sense that the full power of mathematical
>functions is present". IMO, this property is sufficient to classify a
>language as "functional". Even SML does have mutable state (references)
>and assignment, BTW.


So SML is an imperative language and a functional language then.
These days one tends to avoid that, because with mutable state, you
don't get referential transparency, and without referential
transparency, you miss out on some benefits of functional languages;
but in earlier times people did not know how to live without mutable
state, so they had mutable state in functional languages. Quite a lot
in Lisp, less in ML, nothing in Haskell (or at least the imperative
stuff is encapsulated in a way that avoids destroying referential
transparency for the functional part).


- anton
--
M. Anton Ertl
anton@mips.complang.tuwien.ac.at
http://www.complang.tuwien.ac.at/anton/

Return to the comp.compilers page.
Search the comp.compilers archives again.