Sweet ruby logic

Mon, 30 Oct 2006 09:50:52 +0000

One of my on-again, off-again, learning projects over the last couple of months has been the Prolog programming language. It can be a bit of a brain-teaser because unlike languages like Java or Ruby where you write each step of how a program goes about solving some problem, in Prolog you give the program a set of rules and a set of beginning facts and it deduces the solution itself (if it can).

Here's an example Prolog program:

fallible( X ) :- man( X ).
man( socrates ).
?- fallible( socrates ).
yes

In Ruby this might look like:

class Object
    def fallible?
        false
    end
end

class Man
    def fallible?
        true
    end
end

socrates = Man.new
puts socrates.fallbile?

Prolog doesn't have to worry too much about how things get done, it just sort of does them. Once you start grokking it, it's very elegant.

My biggest complaint is the lack of a stylish Prolog environment on MacOSX. I've tried XPG but found it too buggy (in particular it doesn't handle infinite recursion well) and SWI-Prolog. SWI-Prolog is better but it's X and, well, it's just not Mac'ish.

What lies behind my desire to learn Prolog is the ability to build hybrid applications which use Ruby (or maybe Objective-C) where it makes sense, generating declarative Prolog sub-programs to solve problems as necessary.

So I was very interested to read about Mauricio Fernandez experiments with a Ruby implementation called tiny-prolog.

Now the Ruby example might look something like:

fallible[:X] <<= man[:X]
man[:socrates].fact
query fallible[:socrates]

Not quite as nice as the Prolog version but a step forwards none the less.

Curiouser and Curiouser! on prolog

Sweet ruby logic