Comments on: Some simple thoughts on Erlang

By: Rich Seagraves

Rich Seagraves — Tue, 02 Oct 2007 18:49:40 +0000

Actually, Erlang lands somewhere in the middle between Prolog’s unification and pattern matching found in a language like Haskell or (S|OCa)ML. It’s true the variable bindings only go one way (so the variables aren’t Prolog’s Logic Variables), but you can do things like

> {A, A, B} = {abc, abc, foo} %% A binds to abc, B binds to foo

(the above completely and gratuitously copied from http://www.erlang.org/course/sequential_programming.html#patterns)
which you aren’t allowed to do in Haskell or the MLs.

By: Shannon -jj Behrens

Shannon -jj Behrens — Tue, 28 Aug 2007 08:33:43 +0000

Yep, agreed on all counts.

By: Merlyn Albery-Speyer

Merlyn Albery-Speyer — Sat, 25 Aug 2007 17:40:25 +0000

Ted – good summary.

Jay – check out From Python to Ruby on Rails to Erlang
http://slideaware.typepad.com/slideaware/2007/04/from_python_to_.html

By: Jay P

Jay P — Thu, 23 Aug 2007 20:31:20 +0000

I’m really enjoying the book so far. This is the first functional language that I feel like I’ve really understood (maybe that’s just because the book is written so well).

I’m also going to have a hard time finding immediate use for it though. I’m curious as to how Erlang pros do it. Do they use Erlang as their default choice, and move to another language only on rare occasions?

I’m a Python guy, and Python fits so nicely with everything I do, it’ll be hard to find places for Erlang to fit. I’ll certainly try though!

Ted, you mentioned that you knew of some Python/Ruby/Java apps that were mixed with Erlang. Could you mention some? I know that Bob Ippolito does that with MochiAds, but it’s all I’ve heard of. Something open source would be very interesting to see.

By: Ted Leung

Ted Leung — Thu, 23 Aug 2007 20:02:44 +0000

I think that the important applications are yet to come – I’m not sure that it will be that worth it to rewrite existing applications in Erlang unless that application is having perf/scalablity problems.

By: rick

rick — Thu, 23 Aug 2007 18:57:14 +0000

I still haven’t finished the book yet, but one area where it really shows its heritage is its handling of bit-level constructs.
Some of the example code, walking through binary file formats and munging protocol headers is very simple and powerful.
I haven’t really found a good use for the language in my day-to-day work, but if I wanted to analyze network protocol headers, it would be a good choice!
(I wonder if one could think of strings in a similar way, providing language constructs that worked at the character and string level instead of bits/bytes? )

As far as overall language adoption, its not just the syntax but also the semantics. A lot of people don’t “get” pattern-matching and recursive functional style.
It seems like people with Lisp/Scheme/Prolog familiarity pick up on it again quickly (since they do the same thing) but people used only to procedural/OO semantics
have a little more trouble grasping it at first.
Once they do, then its a matter of deciding upon the utility of it!