From rff.rff at gmail.com  Wed Oct 17 09:00:58 2007
From: rff.rff at gmail.com (gabriele renzi)
Date: Wed, 17 Oct 2007 15:00:58 +0200
Subject: [gecoder-users] syntax idea
Message-ID: <828083e70710170600x3efea407mf9b2d62adbe2e832@mail.gmail.com>

Hi everyone,

I just started to play with gecoder few hours ago, and it looks great,
 thanks to all the developers for it.

But I was wondering if an alternate syntax had been thought, or maybe
refused before.
Basically, I was thinking of expressing constraint withouth the .must
and using a subsequent Model#must or (#holds or #post) to post them to
the constraint store.

Kind of more functional and less OO if someone believes that OO just
means .foo :)

So the SENDMOREMONEY example would be:

  def initialize
     s,e,n,d,m,o,r,y = @letters = int_var_array(8, 0..9)
     holds equation_row(s, e, n, d) + equation_row(m, o, r, e) ==
equation_row(m, o, n, e, y)
     holds s > 0
     holds m > 0
     distinct @letters
     branch_on @letters, :variable => :smallest_size, :value => :min
   end

it has the usual problem that != can't be overloaded in ruby 1.8, but
it seems cleaner to me.

I could work out a patch if wanted, but mostly, I was just curious
about others' opinions on this.

From rubyforge at lokorin.org  Wed Oct 17 10:26:00 2007
From: rubyforge at lokorin.org (Andreas Launila)
Date: Wed, 17 Oct 2007 16:26:00 +0200
Subject: [gecoder-users] syntax idea
In-Reply-To: <828083e70710170600x3efea407mf9b2d62adbe2e832@mail.gmail.com>
References: <828083e70710170600x3efea407mf9b2d62adbe2e832@mail.gmail.com>
Message-ID: <47161B78.6010105@lokorin.org>

gabriele renzi wrote:
> But I was wondering if an alternate syntax had been thought, or maybe
> refused before.
> Basically, I was thinking of expressing constraint withouth the .must
> and using a subsequent Model#must or (#holds or #post) to post them to
> the constraint store.
> 
> [...]
> 
> it has the usual problem that != can't be overloaded in ruby 1.8, but
> it seems cleaner to me.
> 

Such a syntax was considered in the initial syntax discussion over at 
ruby-talk[1].

In the end the decision was mainly based on that must/must_not reads 
naturally for most constraints. One of the downsides is that the line 
must be read to understand what is being done, i.e. it's harder to 
quickly scan the lines with must/must_not than with holds.

The syntax is however not nailed down yet and I do encourage a 
discussion about it.

[1] http://blade.nagaokaut.ac.jp/cgi-bin/scat.rb/ruby/ruby-talk/254467

-- 
Andreas Launila