Comments on: Erlang, the functional Smalltalk?

By: links for 2007-12-04 « Mike Does Tech

links for 2007-12-04 « Mike Does Tech — Tue, 04 Dec 2007 00:45:17 +0000

[...] Stifflog - Erlang, the functional Smalltalk? [...]

By: Isaac Gouy

Isaac Gouy — Sat, 01 Dec 2007 21:18:58 +0000

> I think you misunderstood my intentions

I don’t know your intentions, just what you write.

By: stiff

stiff — Sat, 01 Dec 2007 19:28:50 +0000

Pierre, Mark, she: Thanks for the comments

Issac: I think you misunderstood my intentions. I don’t want to prove that Erlang processes are exactly equal to objects in other languages, nor do I want to bring traditional OOP to Erlang, what the mailing list post seems to be about. I just believe that Erlang processes are conceptually very similar to what we intuitively understand as “objects”, and in some ways, ie. by focusing on the cooperation/message-exchange part, they encourage better mapping of real world concepts to “objects” than the mainstream pseudo-OO languages like Java.

By: Pierre R

Pierre R — Sat, 01 Dec 2007 18:15:36 +0000

What Alan Kay said about objects ?

“OOP to me means only messaging, local retention and protection and hiding of state-process, and extreme late-binding of all things”

Another well known and respected equation for objects is (not from me) :
Objects == Lambda Abstraction + Message Dispatch + Local Side Effects

Now there is another model that some OO people seem to ignored completely called the actor model :
Actor == Lambda Abstraction + Message Dispatch (asynchronous between actors) .

These two models are known since the early 70 (and before ).

Erlang obviously does not encourage local side effect and communication between processes is asynchronous.

That is why Erlang is not really perceived as an OO languages …

I hope it helps.

By: Isaac Gouy

Isaac Gouy — Sat, 01 Dec 2007 17:48:15 +0000

Mark Aufflick wrote
> It’s interesting to note that Alan Kay …

It’s also interesting to note how comprehensively what Alan Kay said seems to have been ignored.

By: Isaac Gouy

Isaac Gouy — Sat, 01 Dec 2007 17:46:03 +0000

“The rest of the chapter was a discussion of why we thought that OO *design* might be acceptable but why OO *languages* are not.”

http://www.erlang.org/pipermail/erlang-questions/2007-November/030920.html

> “you have many things in Erlang that aren’t at all realized using processes, whereas in “true” OO languages everything is an object.”

iow sequential Erlang is a functional programming language - nothing about it could be confused with an OO programming language.

> You can also build reusable components with them …

And you can build reusable components in non-OO imperative languages.

> seems much more reasonable than trying to fit the traditional model with classes to the functional world

Have you looked at Scala?

By: she

she — Sat, 01 Dec 2007 15:09:26 +0000

But this is problematic because you can say that a message is in fact an object as well. What I believe will be a good new language is one that easliy combines events and trigger hooks listening on multiple events (i.e. multiple servers), and each “master” object that controls these multiple events can clone itself spawning more events that trigger other events….

I am curious how swarm AI will turn out, in essence I think we could in a few years have very strong interacting robots (real robot teams, not those stupid soccer robots…)

Anyway, I

By: Mark Aufflick

Mark Aufflick — Sat, 01 Dec 2007 12:35:54 +0000

I completely agree. It’s interesting to note that Alan Kay (one of the creators of SmallTalk and OO) says that OO is more about message passing than it is about objects - and Erlang provides a really interesting different angle on message passing.