Testing MIME response types

I feel like I might have covered this before, but I was looking for a way to test respond_to. I had found this post on how to test it, but after looking at it for a while, I found myself rewriting it. Mainly, I took out parts that convert the Mime types, and inserted Rail's own Mime type objects. You can use it like this:

request_mime(:fbml) do 
  get :list
  assert_response_mime(:fbml)
end

request_mime("text/xml") do 
  get :list
  assert_response_mime("text/xml")
end

Just include it in your test_helper.rb file in test/

class Test::Unit::TestCase
  include Threecglabs::MimeTestHelpers
end

Here's "mime_test_helpers.rb". Just throw it in lib/

module Threecglabs
  module MimeTestHelpers

    def self.included(mod)
      mod.class_eval do
        include MimeRequest
        include MimeAssertions
      end
    end

    module MimeRequest
      # changes the mime type of the request within the block
      #
      #   request_mime(:fbml) do 
      #     get :list
      #     assert_response_mime(:fbml)
      #   end
      def request_mime(mime_type_name_or_extension)
        if mime_type_name_or_extension.kind_of?(String)
          mime_type = Mime::Type.lookup(mime_type_name_or_extension)
        elsif mime_type_name_or_extension.kind_of?(Symbol)
          mime_type = Mime::Type.lookup_by_extension(mime_type_name_or_extension.to_s)
        else
          raise ArgumentError.new("mime type must be string or symbol")
        end
        old_mime_type = @request.accepts
        @request.accept = mime_type.to_s
        yield
        @request.accept = old_mime_type
      end      
    end

    # These are assertions to test respond_to, whether they return a specific mime type
    # as a response to a request
    module MimeAssertions

      # Helps out with response testing, by letting to assert that the most recently-made 
      # request responded with a particular MIME extension, like :html, :fbml, :xml, etc.
      def assert_response_mime(expected_mime_type_ext)
        expected_mime_type = Mime::Type.lookup_by_extension(expected_mime_type_ext.to_s)

        # Mime::Type.parse doesn't parse accept parameters correctly, therefore
        # we account for having multiple types in the accept
        response_mime_types = @response.headers['type'].split(/,\s*/).map do |accept_type|
          mime_type_name = accept_type.split(/;\s*/).first
          Mime::Type.parse(mime_type_name)
        end
        assert_block("Responded with #{response_mime_types.map(&:to_s).inspect} when expecting #{expected_mime_type}") {
          response_mime_types.any? { |response_mime_type| expected_mime_type == response_mime_type }
        }
      end

    end
  end
end

Snippet!

irb console tricks

irb Mix Tape — err.the_blog

Not a big entry, but I think worth posting because it pays to learn your tools. I've got no insights of my own to add. Someone else's tip!

Render_to_string only counts if failed

render_to_string followed by render issues - Ruby Forum

I was puzzled by this yesterday, and good thing I went to do something else instead of tearing my hair out. Came back fresh today and found this. Apparently, render_to_string normally doesn't count as a render--given that it succeeds. If it throws an exception for some reason, then it'll get count as a double render if you rendered elsewhere. Small tip!

DRYing up your views with a TableBuilder

In the last two months, when I was adding more features to mobtropolis, I found it painful to try and lay things out all over again from scratch. As a result, it sucked to see ugly layouts on the new pages juxtaposed with all the styling I had done before. It wasn't until a week ago that I said to myself, "Stop the madness!" and started refactoring my views--something I never thought of doing much of until now. When you don't, the barrage of angle brackets blows out of proportion and it starts to look pretty damn fugly with complex views.

What I should be able to do is take common mini-layouts in my views and make them available as helpers so that I can think in terms of bigger chunks to piece together pages, rather than in divs and spans. In addition, it makes your interface more consistent for your users.

Some good resources were presentations from the 2007 RailsConf, like V is for Vexing and Keeping your views DRY. While a lot of view DRYing talks about form builders, I didn't see any on table builders, so I decided to take a stab at it. Personally, I don't like to overuse tables for layouts. But as certain elements in my page layouts have been repeated, I refactored them into first helpers, and then when I did more than one, I extracted it out into a simple table builder. This is how you'd use it:

For example, I have a mini-layout where I show simple stats:


Here's how I used a simple table builder to display the above:

<% score_card do |sc| -%>
  <% sc.placard("Votes", "", num_voters, "") -%>
  <% sc.placard("Sceneshots", "", num_sceneshots, "")-%>
  <% sc.placard("Participants", "", num_participants, "") -%>
  <% sc.placard("Challenges","", num_challenges, "") -%>
<%  end -%>

And I find that I started using the same sort of thing in other places, like in a user's profile:

<% score_card do |sc| -%>
  <% sc.placard("Submitted Scenes", "", num_scenes, "") -%>
  <% sc.placard("Submitted Sceneshots", "", num_sceneshots, "") -%>
<% end -%>

I cut out some details so you can see that it's just a block that gets passed a ScoreCard object, from which you call placard to add another score to the score_card. It sure beats writing <table> and <td> over and over again.

To declare the helper, we create a helper with the structure of the table inside the declaration of a ScoreCard object. We have a ScoreCard object to hold the contents of the placards. When they're called in the block above in the template, they get stored in the ScoreCard object, and not written out to erb immediately. That way, we can place them wherever in the table we please, by making the call to card.display(:placards):

module ScorecardHelper
 def score_card(html_options = {}, &block)
    options = { :class => :scorecard, :width => "100%" }.merge(html_options)
    ScoreCard.new(block) do |xm, card|
      xm.table(options) do
        xm.tr(:valign => :top) do
          xm << card.display(:placards)
        end
      end
    end
  end
end

So then what's ScoreCard look like? Pretty simple. It has a call to each cell that can be filled in the mini-layout. It's kinda analogous to how form_for passes in a form object, on which you can call text_field, etc.

require 'lib/table_builder'

# Used to create a scorecard helper
class ScoreCard < TableBuilder
  cells :placards

  # a placard is placed into scorecards
  def placard(text, text_id, score, widget)
    xm = Builder::XmlMarkup.new
    @placards[:html] += xm.td do
      xm.span(:style => "font-size: 1.2em") { xm << "#{text}" }
      xm.em(:id => text_id, :class => "primary_number") { xm << "#{score}" }
      xm << "#{widget}"
    end
  end

end

Notice that there's a call to cells() to initialize the type of cell, and a method of the same name that builds the html for that cell. If you have other types of cells, you simply put it in the list of cells, and then create a method for it that is called in the template. By convention, you'd stick the html of the cell contents in "@#{name_of_cell}"[:html], and if you wanted, pass in the html_options, and stick it in "@#{name_of_cell}"[:options]. Then, you can access those in the helper wherever you want.

Let's try another one. I have a mini_layout with a picture, and some caption underneath it, like a polaroid.

<% polaroid_card do |p_card| -%>
  <% p_card.picture do -%>
    <%= sceneshot_for scene -%>
  <% end -%>
  <% p_card.caption do -%>
    <%= pluralize(scene.num_of_sceneshots, "sceneshot") %>
  <% end -%>
<% end -%>

The associated helper and PolaroidCard object are pretty simple:

module PolaroidCardHelper
  # a polaroid card is used to show a picture with a caption below it.
  def polaroid_card(html_options = {}, &block)
    options = { :class => :polaroidcard, :style => "display: inline;" }
    PolaroidCard.new(block) do |xm, card|
      xm.table(options) do 
        xm.tr { xm.td(card.html_options(:picture)) {
            xm << card.display(:picture)
          }}
        xm.tr { xm.td(card.html_options(:caption).merge(:class => "caption")) {
            xm << card.display(:caption)
          }}
      end
    end
  end
end

require 'lib/table_builder'

class PolaroidCard < TableBuilder
  cells :picture, :caption

  def picture(html_options = {}, &block)
    @picture[:html] = capture(&block)
    @picture[:options] = html_options
  end

  def caption(html_options = {}, &block)
    @caption[:html] = capture(&block)
    @caption[:options] = html_options
  end
end

I've tried to pull all the plumbing out into TableBuilder (dropped it into lib/), and only leave the flexibility of creating the table structure in the helper, and the format of the cell in the object. It ends up TableBuilder isn't too complex either. It uses some metaprogramming to create some instance variables. I know it pollutes the object instance variable namespace, but I wanted to be able to say @caption[:html], rather than @cells[:caption][:html].

class TableBuilder < ActionView::Base
  class << self
    # used in the child class declaration to name and initialize cells
    def cells(*names)
      define_method(:initialize_cells) do 
        @cell_names = names.map { |n| "@#{n}".to_sym }
        @cell_names.each do |name| 
          if instance_variable_defined?(name)
            raise Exception.new("name clash with ActionView::Base instance variables") 
          end
          instance_variable_set(name, { :html => "", :options => {} })
        end
      end
    end
  end

  def initialize(decor_block, &table_block)
    super
    initialize_cells
    decor_block.call(self)
    html = table_block.call(Builder::XmlMarkup.new, self)
    concat(html, decor_block.binding)    
  end

  def display(cell_name)
    instance_variable_get("@#{cell_name}")[:html]
  end

  def html_options(cell_name)
    instance_variable_get("@#{cell_name}")[:options]
  end

end

I've found have these helpers cleans up my views significantly, though I have to admit, it's not exactly easiest to use yet. In addition, I'm not exactly thrilled about having TableBuilder inherit from ActionView::Base, but it was the only way I could figure out to get the call to concat() to work. In any case, the point is to show you that refactoring your views into helpers is a good idea, and even something like a table builder goes a long way, even if you don't do it the way I did it. Lemme know if this helps or hinders you. snippet!

What the heck is a Ycombinator

I woke up at 4am this morning for no reason and decided to go through Ycombinator in Ruby. Given that I read Hacker News daily, it's a shame I didn't know what a Ycombinator was. I thought the article was pretty clear, at least compared to Raganwald's article. As an aside, from purely a learning experience, it was kinda fun and eye opening. At least I can see why geeks think it's neat.

I had written a reply to a comment on hacker news about it, and it turned out to be really long, so I just extracted it out here. I was suppose to start some coding, but I got wrapped up in writing up the comment. My comment ended up to be pretty long, but it was shorter than it could have been. You guys get to suffer through the long version. Haha. But I'm not going to go into details. You can read that in the linked articles. Instead, I'll paint some broad strokes.

  Y = λf·(λx·f (x x)) (λx·f (x x))

or in Ruby, copied from the aforementioned article:

y = proc { |generator|
    proc { |x|
        proc { |*args|
            generator.call(x.call(x)).call(*args)
        }
    }.call(proc { |x|
        proc { |*args|
            generator.call(x.call(x)).call(*args)
        }
    })
}

or if you prefer elegance, Tom's solution in response to Raganwald

def y(&f)
    lambda { |x| x[x] } [ 
        lambda { |yf| lambda { |*args| f[yf[yf]][*args] } } ]
end

I found lambda calculus above hard to read. However, if you go through the code in Y Combinator in Ruby, you'll find it's not too bad. I find that this lecture is also pretty good, as it takes you step by step, with a little bit of humor as well.

If I had to take a stab at it, A Ycombo is a way to implement recursion mechanism when the language doesn't provide named recursion, loops, or iterators, and all you get are first-class functions and a few substitution rules.

Functions are just mappings from one set of things to another set of things--ie. give it a number, and it'll point to another number. Ycombo relies on a property of functions that sometimes, when you put something into a function, you get the exact same thing out, i.e. something gets mapped to itself, like f(x) = x^2, where f(1) = 1 is an example of this property. They call this the fixed point of a function.

The thing about functions is that they don't just have to map numbers to numbers. They can map functions to other functions. A derivative is an example of a function that takes one function as an input, and spits another function back out. like d(x^2) = 2x. Where is the fixed-point of a derivative? One of them is when you put e^x into it. d(e^x) = e^x. I'm sure there are more.

This is important because if you can find the point in which a function can return a function unchanged, you can use that to call the function again, which is what we call recursion. And all the trickiness you see in ycombinator that you see is mainly a result of functional programming not keeping state, so you have to pass everything you need into a function. So if you have to do recursion, you have to have a mechanism pass the function itself, so you can call it again. And this mechanism kinda bends back onto itself, and that's why you see a part of the ycombinator repeated twice in the code above, and in the lambda calculus.

It seems pretty inane to use ycombo given that modern high level languages provide named recursions, and if anything, for loops and iterators with closures. But what if you don't? How do you process lists/arrays without loops or named recursion? Generally, you'd have to make your own recursion mechanism.

When would you ever have languages that don't have those things? Probably not often. But Forth comes to mind. I've never used Forth, but from what little I know about it, the language starts off with some basic primitives and that's it. No loops, no if statements. How do you get anything done? Well, you build your own control structures. People use Forth because it's possible to build your own very small compiler from the ground up written in Forth itself, and still understand the entire thing. I suspect it's used in embedded programming because of that reason.

So you'd pretty much use it when you want to have a powerful system from first principles. I'm guessing it's possible to build your own computer by implementing only functions and substitution rules in hardware, and then you can derive everything else, numbers, pairings, and even recursions in software. That way, you keep your hardware costs down, while retaining the power of a Turing machine.

Speculating here...but another thing that might be interesting is that ycombinator might be a way to compress code. Software size should be reducible if the code is compressible. And recursion can be thought of as a compressing code, and expanding it when the recursion is run. I wonder if there's other ways to reduce software bloat with the use of a ycombinator besides recursion?

(Aha!) Part of the reason why great hackers are 10 times as productive

I knew in college that some dudes were faster than I was in terms of programming. Since peer programming wasn't exactly encouraged in college, and at work I did mostly prototyping work, I never really knew how fast other programmers worked.

So when I read Paul Graham (and Joel's) claim that great hackers are at least ten times as productive as average programmers (too lazy to cite right now), I was kinda shocked. Surely, ten times is an order of magnitude! Something that takes an average programmer a 40 hour week to do the great hacker can do in a 4 hour afternoon?

I wondered about that, since there are times when I get stuck on something, then I just start stabbing around randomly out of frustration. I had assumed that great hackers were faster only because they had either the experience or insight to side-step whatever I was doing wrong.

But lately, I've been re-reading everyone's essays that write about programming productivity. And one thing that caught my eye the second time around was when Paul Graham was talking about bottom up programming and how he didn't really believe in objects, but rather, he bent the language to his will. He was building new blocks for himself so he could think about the problem at a higher level of abstraction.

This is basic stuff! I mean, that's the whole point of higher-level programming. When you refactor methods out, you're creating a vernacular so that you can express the problem in terms of the problem domain, not in terms of general computing. This is much like if you want to drive a car, you'd want to be able to step on the gas, rather than time the firings of the pistons. And if you want to control traffic in a city, you'd rather tell all cars to go to a destination, rather than stepping on the gas and turning the steering wheel for each one.

But taken into the light of bending a language to your will, it makes it more clear for me as to how great hackers are ten times as productive. Great hackers are productive not only because they know what problems to sidestep and can problem solve systematically and quickly, but they also build a set of tools for the problem domain as they go along. They are very good pattern recognizers and will be able to generalize a particular pattern of code, so that they can use it again. But not only that, great hackers will create an implicit understanding attached to the abstraction, ie. what we might call common sense.

A case in point. Before Ruby, I'd used for loops over and over again, never really thinking that I could abstract a for loop. It wasn't until they were taken away in Ruby did I realize that map, inject, and their cousins are all abstractions of the for loop. When I see "map" I know that it performs a transformation on every element. But I also know that the array I get back will be the same size, that each element operation doesn't affect other elements, among other things. These are implicitly stated, and they allow for shorter code.

When that happens, you can simply read "map", and get all the connotations it comes with, and hence it comes with meaning. It also becomes easier to remember, since it's a generalized concept that you can apply in different places in the code. The more times you use it, the easier it is to remember, instead of having specialized cases of the same kind of code where the behavior is different in different parts of the code.

A great hacker will take the initial time upfront to create this generalized code, and will save in the long run being able to use it. Done over and over again, it all adds up. So it's not that for any given problem, a great hacker will be done in 4 hours what it takes an average programmer 40 hours, but that over time, a great hacker will invest the time to create a tools and vocabulary that lets him express things easier. That leads to substantial savings in time over the long haul.

I hesitated writing about it, as it's nothing I (nor you) haven't heard before. But I noticed that until recently, I almost never lifted my level abstraction beyond what the library gave me. I would always be programming at the level of the framework, not at the level of the domain. It wasn't until I started writing plugins for rails extracted from my own work and reading the Paul Graham article that a light went off for me. It was easier to plug things like act_as_votable together, rather than to still mess around with associations (at the level of the framework). I still believe you should know how things work underneath the hood, but afterwards, but all means, go up to the level of abstraction appropriate for the problem domain.

DSLs (Domain specific languages) are really just tool-making and vernacular creation taken to the level of bending the language itself. It's especially powerful if you can add implicit meaning to the vernacular in your DSL. It's not only a way of giving your client power in their expression, but it's also a refactoring tool so that you can better express your problem in the language of the problem domain. Instead of only adding methods to your vernacular, you can change how the language works. It was with this in mind that I did a talk on DSLs this past weekend at the local Ruby meetup. First part is on Dwemthy's Array, and the second is using pattern matching to parse logo. Both seemed pretty neat when I first read about it. Enjoy!

---

DSL in Ruby through metaprogramming and pattern matching

Generating rdoc for gems that refuse to generate docs

I recently upgraded to capistrano 2.1, and it's woefully lacking in documentation. Jamis Buck had already picked a documentation coordinator about a month ago, but nothing seemed to be happening since.

So it was time to go dumpster diving in cappy code. I at least wanted to see what the standard variables were. To my surprise, there were some docs in code, but I couldn't generate it with

gem rdoc capistrano

For those of us that had never made a gem, all you have to do to force it to is to edit the associated specification for the gem (/usr/lib/ruby/gems/1.8/specifications/), and add

s.has_rdoc = true

Maybe if I dig enough stuff out of it, I'll pose some prelim documentation for cappy 2.1 and donate it to the documentation effort.

As an aside and musing, ideally, the code itself would be documentation. However, just because you're reading code, doesn't mean that you can get the overall picture of how to use it. Even if you filtered out the details, and saw just the class and method declarations, that still wouldn't be enough since you can't see how things fit together. I don't know what a good solution would be. The simplest API is often complex enough to be nearly useless without good documentation.

Small tip!

The user owns this and that

In most any community-based website, your users are creating objects. And sometimes, you'd only like to display them when they're owned by the user. Usually, it's easy enough to do the check based on the association in the template views.

  <% if @story.account == session[:member] %>
    <%= render :partial => "story"
  <% end %>

However, for the sake of perhaps over-reaching, but something more readable, we try:

class Story < ActiveRecord::Base
  belongs_to :account

  def owned_by?(user)
    self.account == user
  end
end

  <% if @story.owned_by?(session[:member]) %>
    <%= render :partial => "story"
  <% end %>

But this gets to suck, because you have duplicate code in different classes that are just slightly different due to association names. One way to solve it is to put it in a module, and include it in different classes. After all, that's what mixins are for.

module Ownership
  def owned_by?(user)
    self.account == user
  end
end

class Story < ActiveRecord::Base
  include Ownership
  belongs_to :account
  # blah blah blah
end

class Post < ActiveRecord::Base
  include Ownership
  belongs_to :account
end

Or alternatively, you can put it in the account, but in reverse. But now, you have to search through the associations of the owned ActiveRecord object.

class Account < ActiveRecord::Base
  def owns?(ar_object)
    ar_object.class.reflect_on_all_associations(:belongs_to).detect { |assoc|
      ar_object.send(assoc.name) == self
    } ? true : false
  end
end

I find the tertiary operator to be kinda ugly there, but it doesn't make sense to return the reflection object. Regardless, this lets you do:

  <% if session[:member].owns?(@story) %>
    <%= render :partial => "story"
  <% end %>

However, this doesn't work for self-associated AR objects, or objects that have two or more belongs_to to the same account. It relies on a unique belongs_to association for every object belonging to account. I'm not sure yet, which way's the better way to go, and in the end it probably doesn't matter as much, but I do like being able to say user.owns?(anything) for any object without really thinking about what the associations are. half-tip.

A simple distributed crawler in Ruby

A week ago, I took a break from Mobtropolis, and...of all things ended up writing a simple distributed crawler in Ruby. I hesitated posting it at first, since crawlers are conceptually pretty simple. But eh, what the heck.

This was just an exercise to do some DRb and Hpricot, so don't use this for your production work, whatever it may be. An actual crawler is far more robust than what I wrote. And don't keep it running hammering at stuff, since it'll get you banned.

First, this is how you use it:

WebCrawler.start("http://en.wikipedia.org/") do |doc|
  puts "#{doc.search("title").inner_html}"
end

And that's it. It returns documents in an XPath traversable form, courtesy of Hpricot.

A web crawler is a program that simply downloads pages, takes notes of what links there are on that page, and puts those links on its queue of links to crawl. Then it takes the next link off its queue and downloads that page and does the same thing. Rise and Repeat.

First, we create a class method named start that creates an instance of a webcrawler and then starts it. Of course, we could have done without this helper method, but it makes it easier to call.

module Crawler
  class WebCrawler
    class << self
      def start(url)
        crawler = WebCrawler.new
        crawler.start(url) do |doc|
          yield doc
        end
        return crawler
      end
    end
end

So next, we define the initialization method.

module Crawler
  class WebCrawler
    def initialize
      puts "Starting WebCrawler..."
      begin
        DRb.start_service "druby://localhost:9999"
        puts "Initializing first crawler"
        puts "Starting RingServer..."
        Rinda::RingServer.new(Rinda::TupleSpace.new)
        
        puts "Starting URL work queue"
        @work_provider = Rinda::RingProvider.new(:urls_to_crawl, Rinda::TupleSpace.new, "Queue of URLs to crawl")
        @work_provider.provide
        
        puts "Starting URL visited tuple"
        @visited_provider = Rinda::RingProvider.new(:urls_status, Hash.new, "Tuplespace of URLs visited")
        @visited_provider.provide
      rescue Errno::EADDRINUSE => e
        puts "Initialize other crawlers"
        DRb.start_service
      end
      puts "Looking for RingServer..."
      @ring_server = Rinda::RingFinger.primary
      
      @urls_to_crawl = @ring_server.read([:name, :urls_to_crawl, nil, nil])[2]
      @urls_status = @ring_server.read([:name, :urls_status, nil, nil])[2]
      @delay = 1
    end
  end
end

This bears a little explaining. The first webcrawler you start will create a DRb server if it doesn't already exist and do the setup. Then, every subsequent webcrawler it'll connect to the server and start picking URLs off the work queue.

So when you start a DRb server, you call start_server with a URI, then you start a RingServer. What a RingServer provides is a way from subsequent clients to find services provided by the server or other clients.

Next, we register a URL work queue and a URLs visited hash as services. The URL work queue is a TupleSpace. If you haven't heard of TupleSpace, the easiest way to think of it is as like a bulletin board. Clients post items on there, and other clients can take them out. This is what we'll use as a work queue of URLs to crawl.

The URLs visited is a Hash so we can check which URLs we've already visited. Ideally, we'd use the URL work queue, but DRb seems to only provide blocking calls for reading/taking from the TupleSpace. That doesn't make sense, but I couldn't find a call that day. Lemme know if I'm wrong.

module Crawler
  class WebCrawler
    def start(start_url)
      @urls_to_crawl.write([:url, URI(start_url)])
      crawl do |doc|
        yield doc
      end
    end

    private

    def crawl
      loop do
        url = @urls_to_crawl.take([:url, nil])[1]
        @urls_status[url.to_s] = true
        
        doc = download_resource(url) do |file|
          Hpricot(file)
        end or next
        yield doc

        time_begin = Time.now
        add_new_urls(extract_urls(doc, url))
        puts "Elapsed: #{Time.now - time_begin}"
      end
    end
  end
end

Here is the guts of the crawler. It loops forever taking a url off the work queue using take(). It looks for a pattern in the TupleSpace, and finds the first one that matches. Then, we mark it as 'visited' in @urls_status. Then, we download the resource at the url and use Hpricot to parse it into a document then yield it. If we can't download it for whatever reason, then we grab the next URL. Lastly, extract all the urls in the document and add it to the work queue TupleSpace. Then we do it again.

The private methods download_resource(), extract_urls(), and add_new_urls() are merely details, and I won't go over it. But if you want to check it out, you can download the entire file. There are weaknesses to it that I haven't solved, of course. If the first client goes down, everyone goes down. Also, there's no central place to put the processing done by the clients. But like lazy textbook writers, I'll say I'll leave that as an exercise for the readers. snippet!

---

webcrawler.rb

Communicating your intent in Ruby

I've been using Ruby most everyday for about two years now. While I'm no expert, I know enough to be fairly productive in it. And beyond liking the succinctness and power that you often hear other people talk about, it's made me a better programmer. But there's an aspect of Ruby that worries me somewhat.

To start, programming is recognized rightfully as a means to build something from pure thought. But it's also a form of communication, to other programmers that will touch your code later, and to yourself when you look at it months from now. We're at a point that other than embedded and spacecraft programming, we have the luxury of using programming languages that focus ease for the programmer, rather than for the ease of the machine. Fundamentally, that's the philosophy that Ruby takes.

And while Ruby's nice in a lot of ways, I'm not sure about how it communicates an object's interface. When you're allowed to modify objects and classes on the fly, how do you communicate interfaces between modules you mixin and methods/modules you add? By interface, I mean, how do you use this class so that it does what it's suppose to? Normally, it's pretty obvious--you look at the names of the methods declared in the code. A well-written class has public methods exposed, or you look at its ancestor's public methods. You might need some documentation to figure out how to call them in the right order, but generally, you have some idea just by looking at the method signatures.

However, when you throw mixins and metaprogramming in the mix, it becomes less easy to tell just from looking at the method signatures in the code--the structure of the code. You have to specifically read the code, or you have to rely on someone who knew intent to document it in detail.

An example communicating interfaces for mixins: the module Enumerable contains a lot of the Collections related methods. The cool thing is that if you wanted these functions in your own class, all you have to do is define each() in your class, mixin the Enumerable module, and you get all of these "for free". However, outside of documentation explicitly stating it, it's not as immediately obvious in method signatures that this is what you have to do in order to use it. It's only after scanning through the entire code that you notice each() being used for all the methods.

Of course, Ruby contains enough metaprogramming power to protect yourself against this. one can do something like this:

class MethodNeededError < RuntimeError
  def initialize(method_symbol, klass)
    super "Method #{method_symbol.to_s} needs to be in client class #{klass.inspect}"
  end
end

module Enumerable
  def self.included(mod)
    raise MethodNeededError.new(:each, mod) unless mod.method_defined?(:each)
  end
end

This only works if you put the include after you define each(). That's just asking for trouble when the order of your definitions in your class matter.

A fair number of people are writing mini-DSLs in ruby using metaprogramming tricks. One of the common ones is the use method_missing to define or execute methods on the fly. ActiveRecord's dynamic finds are implemented this way. The advantage of communication of interface here in the structure of the code is obvious. Unless it was documented well, you can't tell just by looking at the method signatures.

Why do I harp on interface signatures? I mean, in the instance of requiring each(), it works by just letting it fail in the enumerated methods, since it'll complain about each itself. In the instance of method_missing, just read the regex in the body. While these are true, none of these allow for rdoc to generate proper documentation. The whole point of documentation is to show you the interface--how to use that piece of code. I'm just afraid that given Ruby's philosophy of being able to write clear, powerful, and succinct code, it might fall short when people start using these metaprogramming tricks like alias_method_chain and method_missing more and more. Maybe rdoc needs to be more powerful and read code bodies for regex in method_missing?. It already documents yields in code bodies, but that seems awfully specific.

I'm not a exactly a fan of dictating interfaces like in Java. When you're first coding something up, you're sketching, so things are bound to change. Having plumbing like interface declaration gets in the way, imo. However, when something's a bit more nailed down, it'd be nice to be able to communicate to other programmers your intent without them having to read code bodies all the time.

In the end, I side on flexibility. However, I kinda wish Ruby had some type of pattern matching for methods so I didn't have to read method_missing all the time. But then again, that would be messy in all but the simplest schemes. Can you imagine a class that responded to email addresses as method calls? I guess I'd have to file this one under "bad ideas"

Don't reopen ActiveRecord in another file

The power of Ruby lies partially in how one can reopen classes to redefine them. Besides namespace clashes, this is usually a good way to extend and refine classes to your own uses. However, last night, I got bitten in the ass trying to refactor a couple classes. In Rails, you're allowed to extend associations by adding a class the association call.

class User < ActiveRecord::Base
  has_many :stories, :through => :entries, :source => :story, 
    :extend => StoryAssociationExtensions
end

where StoryAssociationExtensions is a class module holding methods, like expired() that I can perform on the challenges association, so I can do stuff like

  @user = User.find(:first)
  @user.stories.expired  # gives all expired stories

So when refactoring and cleaning up, I renamed StoryAssociationExtensions to AssociationExtensions and reopened up Story class and put it in there. I just wanted to clean up the namespace, and put the association extensions somewhere that made semantic sense. Naturally, I thought putting association extensions for a class belongs in a class. Well, it doesn't work. And don't do it. Hopefully, I'm saving you some pain.

class Story < ActiveRecord::Base
  module AssociationExtensions
    def expired
      self.select { |c| c.expired? }
    end
  end
end

Well, this works if you've reopened the class within the same model file, story.rb in this case. However, if you reopen the class in another file elsewhere, your model definition won't get loaded properly, which leads to associations and methods you defined not to exist.

So imagine my bewilderment when associations didn't work on only certain ActiveRecord Models. In addition, they worked on the unit tests and script/console, but didn't work when the server was running. All that at 3am in the morning. :(

Good thing for source control, so I could revert (but I have to say, svn isn't as easy to use as it could be).

I ended up creating a directory in model called collection_associations and putting the associations in there under a module CollectionAssociations namespace. Not exactly the best arrangement but it'll do for now.

I'm still not sure why ActiveRecord::Base instances don't like being reopened, but I'm guessing it has something to do with only getting loaded once. If anyone has an explanation, I'd like to read about it.

free warning!

State change observer for ActiveRecord

When I started writing some code recently, I noticed that my controllers were getting fat. There was much to do, but there was a bunch of stuff in there that didn't have anything to do with actually carrying out the action--things like sending notifications. ActiveRecord already has observers to take action on certain callbacks. However, what I needed was to take actions on certain state transitions. Not seeing any immediate solutions in the Rails API, I decided to test myself and try writing one. I was bored too. So while I'm not sure if it was worth the time writing it, it certainly was kinda interesting. Here's what I came up with:

Just as a contrived example, let's say we are modeling the transmission of a car. It has three modes: "park", "reverse", "drive". We want to send a notification when a user tries to change it from "reverse" to "drive", but not when he tries to change it from "park" to "drive". If it didn't matter, and we just wanted to send notifications when the state changed to drive, we'd just use the observers that came with ActiveRecord. But since we do care where the state transition came from, here's what I came up with:

class CreateCarTransmission < ActiveRecord::Migration
  def self.up
    create_table :car_transmission do |t|
      t.column :engine_id, :integer, :null => false
      t.column :mode, :string, :null => false, :default => "park"
    end
  end

  def self.down
    drop_table :car_transmission
  end
end

class CarTransmission < ActiveRecord::Base
  include StateTransition::Observable
  state_observable CarTransmissionNotifier, :state_name => :mode
end

So then for my notifier I have:

class CarTransmissionNotifier < StateTransition::Observer
  def mode_from_drive_to_reverse(transmission)
    # send out mail and flash lights about how this is bad.
  end
end

And that's it. Whenever in the controller, I change the state from "reverse" to "drive", lights will flash and emails will be sent out condemning the action, and my controllers stay small and lean.

class CarController < ApplicationController
  def dismantle
    @car = Car.find(params[:id])
    @car.update_attribute :mode, "reverse"
    @car.update_attribute :mode, "drive"
  end
end

So where's the magic? It took a bit of digging around. There were two major things I had to do. I had to insert observers during initialization and I had to override setting of attributes to include an update to notify observers.

ActiveRecord doesn't exactly allow you to override the constructor. I don't think I tried too hard to mess around with it. Looking on the web, I happened upon has_many :through again, where he has some good tips that helped me through Rail's rough edges. While I didn't exactly follow his advice, I did find out about the call back, :after_initialize. It must be something new, because I don't see it in the 2nd edition of the Rails book, and the current official API doesn't list it. Other Rails API manuals seem to be more comprehensive, like RailsBrain and Rails Manual.

Then overridding attributes has always been a bit of a mystery. I found a listing of the attribute update semantics, which was helpful to figure out what I was looking for, but it was false, in that you can't use the first one (article.attributes[:attr_name]=value) to set an attribute. Looking in the Rails code for 1.2.3, it shows that attributes is a read_only hash. But it's right that you should override the second one (article.attr_name=value), since update_attribute() and update_attributes() depends on it.

Again, it ends up that the function I was looking for wasn't found in the official API as a method, other than a short mention in the description of ActiveRecord under Overriding Attributes, which makes it harder to find. Ends up that we can use write_attribute().

So that's pretty much it. Using some standard meta-programming like how plugins do it, you wrap it up, and it's pretty simple:

require 'observer'

module StateTransition
  module Observable
    class StateNameNotFoundError < RuntimeError
      def message
        "option :state_name needs to be set to the name of an attribute"
      end
    end

    def self.included(mod)
      mod.extend(ClassMethods)
    end

    module ClassMethods

      def state_observable(observer_class, options)
        raise StateNameNotFoundError.new if options[:state_name].nil?
        state_name = options[:state_name].to_s
        
        include Object::Observable

        define_method(:after_initialize) do 
          add_observer(observer_class.new)
        end

        define_method("#{state_name}=") do |new_state|
          old_state = read_attribute(state_name)
          if old_state != new_state
            write_attribute(state_name, new_state) # TODO yield the update method
            changed
            notify_observers(self, state_name, old_state, new_state)
          end
        end
      end

    end
    
  end

  class Observer
    def update(observable, state_name, old_state, new_state)
      send("#{state_name}_from_#{old_state}_to_#{new_state}", observable)
    rescue NoMethodError => e
      # ignore any methods not found here
    end  
  end

end

I had a difficult time figuring out how to define methods for an instance of a class. The only thing I came up with was to use define_method, or to include a module with instance methods in them. instance_eval() didn't work. The meta programming for ruby gets rather confusing when you're doing it inside a method--it seems hard to keep track of which context you're in.

So if you can make a use of this, great. If you think it's worth moving it into a plugin, let me know that too. If you know of a better way, by all means, let me know. tip!

Where do you put the rules of Monopoly?

It's apparently a favorite interview question. I have to admit, the first response in my head wasn't a great one, which was "everywhere".

What are game rules? When browsing the rules of popular games like Monopoly and Scrabble, they seem to follow a similar format:

• The initial conditions of the game (the setup)
  
• Then given a condition,
• the set of allowable actions for the player to do
  
• the effects of the condition
  

I embarked on somewhat of the same problem, though on a much smaller scale. I wanted to implement a point system that gave points to those that participated in a Rails app. My first reaction was to make a Rule abstract class, and have different rule classes that subclass it. Something like this (variable names have been changed to protect the innocent):

class SceneRules
  def initialize(model)
    @scene = model
    @user = @scene.submitting_user
  end

  def on_submit
    @user.points += 1
  end
end

class SceneshotRules
  def initialize(model)
    @sceneshot = model
    @user = @sceneshot.sceneshot_uploader
  end
 
  def on_submit
    @user.points += 10
    @scene.submitting_user.points += 5
  end

end

Then I'd be able to call it from the controllers. However, on second thought, it's rather ugly, since I'd be updating the karma everywhere in the controllers. If I understand cross-cutting concerns correctly, scoring karma would be a good example of one. I suppose it's a good candidate for aspect orientated programming, so I scrapped the code above.

Logging is often cited as the poster-child problem to solve with AOP. Logging needs to be done everywhere in the code, but it really has nothing to do with the responsibilities of the class that it's performed in. So you have the same code doing the same thing, duplicated everywhere because there's no one place to put it to make things easy to change.

By the same token, game rules and scoring are of the same nature. And because game rules involve lots of different objects at once, and scoring is interspersed throughout, I think that makes it a good candidate for AOP. However, Ruby has no such direct support for AOP. Instead, the closest thing we have are observers, before/after/around filters (in Rails), and some meta-programming.

I wanted something that allowed me to list out rules like games like Monopoly and Scrabble. I'd have a setup, and some conditions and their effects. Scoring is simplified here because the only time you can score is when one of the models is created or changes state. This is a good fit to the observers and the filters available in Rails.

class ScoringRules < ActiveRecord::Observer
  observe Sceneshot, Scene
  include Rules   

  setup { :scene => Proc.new { |sceneshot| sceneshot.scene } },
        { :sceneshot_uploader     => Proc.new { |sceneshot| sceneshot.sceneshot_uploader },
          :scene_submitting_user => Proc.new { |sceneshot| sceneshot.scene.submitting_user } }

  rule :after_create_sceneshot do |board, players|
    players[:scene_submitting_user].score += 5
  end

  # put more rules here

  def after_create(model)
    rule_dispatch(:after_create, model)
  end

end

Here, the Rules module is what encapsulates the setup, rule, and rule_dispatch calls. I needed setup so that I can access different "game elements" (the board and the players) to update the scoring. It basically stores the setup as a list of lambdas that it can execute at a later time when the rule needs to be executed. Now, when a model is created, we ask the rule dispatcher to figure out which rules execute based on the rules we've named, and then execute the attached block. The block is passed a hash of different game pieces that it needs to update the score and the game conditions. That's it.

I thought it was an interesting way about it and probably warranted some criticism. Is there any particular disadvantage of doing it this way? And if you can think of a way to not have to explicitly state the model relationships in the setup, that'd be nice. half-tip!

Diving into Rails source and explaining alias_method_chain in pictures

I've been involved with the local Ruby group, and I did a talk on diving into the Rails source code. I have to admit, I probably should have dove into it earlier, but there was a lot to learn just by using Rails, for someone that was doing engineering work before this. Originally, I was going to just relate my experience of when I had to dive into the code, but then I happened on Rails' Unusual Architecture, and it gave a good topic for introducing metaprogramming used in Rails.

At first, it was pretty confusing, since I came from a C++ background and was use to the idea of design patterns. Well, there are patterns, it's just that in dynamic languages, much of the traditional Gang of Four design patterns goes away.

I'll outline the meat of the talk. Rails doesn't use the Decorator Pattern. Instead it basically renames methods in calls to inject additional functionality that wraps around the original functionality. How does it do this? It uses something called alias_method_chain, detailed here and here.

And inside alias_method_chain, is a native method called alias_method. Since everything in Ruby is an object, and changeable, that means methods of modules and objects are also changeable. It basically makes a copy of the old method and calls it something new.

I made pictures to show the progression. Let's say we have a method called "save" that we want to enhance with "validation".


We start with ActiveRecord and the Validation module. So in our validation module, we make a method called "save_with_validation"(purple heart) that calls a method called "save_without_validation" inside it. "save_without_validation" doesn't exist yet, because we haven't called alias_method_chain().


Then our first step is to include the Validation module inside of ActiveRecord.

Then when we call alias_method_chain(:save, :validation), using alias_method, it'll make a copy of the original save method, and call it "save_without_validation".


Then, it'll rename the method inside of validation from "save_with_validation" to "save". This way, any code calling save() on ActiveRecord will execute the new save, which does validation first, and then turns around to call the original save (green sun). The client won't know any different, but in fact, code was injected between the original save (green sun), and the caller inside of the new "save" (purple heart). And the original save doesn't know any different either, since nothing's changed from its point of view.

In code, it'll look something like this:

class ActiveRecord
  def save
    # do saving stuff
  end
end

module Validation
  def save_with_validation
    # do validation
    save_without_validation
  end
end

class ActiveRecord
  include Validation
  alias_method_chain :save, :validation
end

Which if we 'executed' the above as we did in pictures, it ends up being equivalent to:

class ActiveRecord
  def save_without_validation
    # do saving stuff
  end

  def save
    # do validation
    save_without_validation
  end
end

I'm not sure how I feel about it as of yet, but at first glance, it's a nice pattern once you know what's going on. At first, I saw all these methods being called in the Rails source which aren't actually defined anywhere. It ends up it's because of metaprogramming stuff like this going on. I think other people have said this is bad idea because it's hard to inject functionality in the middle of chains that already exist. If you pick up one, you pick up everything before it. One might argue the same is true of Decorator patterns.

In any case, you'll see this repeated over and over again in the Rails source, so hopefully, this'll give you some idea of what's going on if you ever decide to go Rails splunking.

Syntactic sugar for dealing with empty containers

In any web application, we're often just reading a collection of rows from the database and displaying it in the browser. Often times, we'll have code that looks like this:

<% unless @friends.empty? -%>
  <% @friends.each do |friend| -%>
    <li><%= h friend.username %></li>
  <% end -%>
<% else -%>
  No friends yet
<% end -%>

I don't know why, but this kinda gets to me, and doesn't look all that neat. I probably attribute it to having to upgrade and maintain a piece of C server code that was nested 8 or 9 layers deep all in one huge main(). It might be counter-productive, but I tried to see if I could do better.

<% if @friends.each do |friend| -%>
  <li><%= h friend.username %></li>
<% end.empty? -%>
  No friends yet
<% end -%>

Well, this is kinda nice in a way that it's only one hierarchy deep. When I look at it, one section is for what to display when there are elements in the list, and one is for when there isn't. I suppose your mileage may vary. However, I didn't like the "if" in front. It obscures the intent of displaying the list. So, in the pursuit of more counter-productivity and perhaps in the spirit of pseudo-altering the language, I tried this out:

<% @friends.each do |friend| -%>
  <li><%= h friend.username %></li>
<% end.empty do -%>
  No friends yet
<% end -%>

Well, that worked. I kinda like it. Since the message was so simple, I had wanted empty() to take a message, and just display it, but because it's a "%" and not a "%=", the message won't get displayed, so I had to do it in a block. In a way, it's almost like being able to write my own "else" statement. If I had used curly braces instead of "do/end", it might look pretty close. Here's the code for empty:

class Array
  def empty(message = "")
    if self.empty?
      return block_given? ? (yield message) : message
    end
  end
end

Like it? Hate it? Tip!

Using anonymous functions inside functions

I was reading an article on Javascript and its use of first class functions. I'd quote it if I could find it again, but basically, it extolled the virtues of being able to create anonymous functions, blocks, and closures easily in Javascript. One of the things that it claimed was that creating local functions only used within a method makes the code much more readable. I was skeptical, but as I was refactoring this function, I thought I'd give it a shot even though I was coding in Ruby. Ruby also allows for the creation of first class functions.

It's just a helper function that generates a bookmark button--however, what gets generated depends on whether a user is logged in and whether a user had already bookmarked the page.

  def bookmark_button(page, user, indicator_id)
    if page.bookmarked_by? session[:user]
      "Bookmarked!"
    elsif page.done_by? session[:user]
      link_to_remote image_tag('heart_add.png', :height => 16, :width => 16) + " Bookmark again!", 
                               :url => { :controller => :bookmark_list, :action => :add, :page_id => page.id  }, 
                               :loading => "Element.toggle('#{indicator_id}')", 
                               :complete => "Element.toggle('#{indicator_id}')"
    else
      link_to_remote image_tag('heart_add.png', :height => 16, :width => 16) + " Bookmark!", 
                               :url => { :controller => :bookmark_list, :action => :add, :page_id => page.id  }, 
                               :loading => "Element.toggle('#{indicator_id}')", 
                               :complete => "Element.toggle('#{indicator_id}')"
    end
  end

Well, I thought it was pretty straightforward to read, though half of it was duplicated code. It wasn't very DRY. So instead of fudging around with another way to structure the control flow, I tried my hand at using blocks. I ended up with this:

  def bookmark_button(page, user, indicator_id)
    bookmark_link = Proc.new { |label| 
      link_to_remote "#{image_tag('heart_add.png', :height => 16, :width => 16)} #{label}", 
                     :url => { :controller => :bookmark_list, 
                               :action => :add, 
                               :page_id => page.id  }, 
                     :loading => "Element.toggle('#{indicator_id}')", 
                     :complete => "Element.toggle('#{indicator_id}')"
    }

    if page.bookmarked_by? session[:user]
      "Doing this"
    elsif page.done_by? session[:user]
      bookmark_link.call("Bookmark again!")
    else
      bookmark_link.call("Bookmark!")
    end
  end

As you can see, I put the link_to_remote() code into a block, since the only difference between the two is the label. In addition, it is possible for the block to use variables that are in this method, but outside the block, even if the execution is outside the scope of this function, if the block is ever passed outside of bookmark_button. I know in javascript, this causes memory leaks in IE. I haven't heard that it's a problem in Ruby yet.

But in this case, I'm just creating a function inside a method to make things more readable and cleaner. I think for the most part, it worked pretty well. As long as it's not overdone (notice I didn't create a proc for the labels), it should make for more readable code, and you don't pollute the class namespace with methods that are only used in this one method. I should get use to this. tip!

Parallel array processing with zip()

When I'm working in Ruby, it's a blast, but a lot of the concepts are borrowed from functional style programming, and I'm not sure what things are called. Often times, I end up writing the same boiler plate code over and over again, refactoring it into a function that I add to the standard library on my own, while knowing there's probably a built-in function to do that. It was like that with inject(). Today, I discovered zip().

I run into this particular problem when I want to line up arrays with one another, and perform an operation on the same corresponding element of each array--much like a parallel map function. For example, vector addition and multiplication is like this. When you add arrays, you take the the nth element from each vector and add them together.

Usually, I write something like this:

a,b = [1,2,3], [2,3,4]
class Array
  def parallel_map(b)
    result = []
    each_with_index do |e, i|
      result << yield e, b[i]
    end
  end
end

a.parallel_map(b) { |ea, eb| ea + eb }

Which is pretty ugly, as I shamefully admit. However, due to the post I last mentioned about Erlang, as well as this post on Ruby talk, I (finally) made the connection that there's such a function in ruby, and it's (also) called zip().

The solution then would look something like this:

a.zip(b).map { |i, j| i + j }

MUCH better. That means something like dot_product is a lot more elegant now.

def dot_prod(a, b)
  a.zip(b).map { |i, j| i + j }.inject { |t, e| t += e }
end

To be honest, I've read the reference API on Ruby's enumerable, array, and hash libraries before. However, at the time that I read zip(), I was thinking, "I'm not sure when I'll ever need to do that.", and just put it out of my mind. :P
tip!

Using ensure in yielding methods

I never really ever find too many occasion to use 'ensure'. It's a Ruby keyword that you can use for blocks of code that ensures, no matter what happens, exceptions or not, the code will be run when the block is finished. And then a quickie that I found in the rails core:

  def silence_warnings
    old_verbose, $VERBOSE = $VERBOSE, nil
    yield
  ensure
    $VERBOSE = old_verbose
  end

It does something simple, just silences the warnings for a particular block of code. On first glance, I would have just written it without the 'ensure'. However, that won't work for yielded blocks that call return or if exceptions are thrown in it, I think. This way, no matter what happens in the block, it will always restore the state that it changed.

Erlectricity: Hi Ruby, I'm Erlang.

Educate. Liberate. - Erlectricity: Hi Ruby, I'm Erlang.People are thinking along the same lines I am, though I have nowhere near the guts at the moment to write a bridger...unless this one sucks.

I think lots of people are excited about the language, though I think that those that are just finding Ruby probably won't look at or touch it for a while longer. "Why do I have to learn yet ANOTHER programming language?" They ask.

In any case, due to Erlang's structure around the Actor Model, it makes algorithms like particle swarm optimization and ant optimization seem pretty exciting. More to come later. In the meantime, enjoy the bridge.

Ruby Snippet: Caching object attributes

Here's another little snippet. I'm not sure if I'm extracting out boiler plate code prematurely, but it seems right.

A good general rule of thumb for refactoring is that one should always call an object for its value, rather than storing object values in temporary variables when you're using the object. Generally, you can get away with setting the value of an object to a temporary variable, and then using that temporary variables for subsequent calculations

# c = Collection.new((0..500).to_a)
collection_sum = c.sum
collection_sum * 3 + 2 # some calculation with the sum
if collection_sum == 4 # some comparison with the sum.
  # do something else
end

You're essentially caching the value outside of the object, 'c'. This is good if 'sum' is an expensive call. However, if the code segment gets large, this method tends to get confusing, especially if you stored it in temporary variables at different times, etc. (Also a good rule of thumb is that your scope should never be bigger than about 10-20 lines) It's better to refer to the value directly from the object.

# c = Collection.new((0..500).to_a)
c.sum * 3 + 2 # some calculation with the sum
if c.sum == 4 # some comparison with the sum.
  # do something else
end

This is not applicable to functional programming languages, since all variables are consts within a scope. But for object-orientated imperative languages, this generally makes for less messy code.

However, performance is a problem, if you have to calculate the sum each time you needed it. An easy solution would be to cache the value. An example would be summing the values across some collection. Normally, that wouldn't be a problem, except if you had hundreds thousand entries. If nothing was added to the collection, the sum would stay the same. If something was added, mark the collection to need updating, and then the next time sum() is called, update the value.

You'd need both a variable to keep track of whether the attribute needed updating, and another variable to hold the result. If you only had one attribute, that'd be ok. If you have more than one, it starts to get a little bit messy. This is where meta programming comes in. I wrote a piece of code that did the bookkeeping of caching attributes for you. You'd use it like:

class Collection
  include AttributeCache

  def initialize
    @array = []
    cache :sum, :initial => 0
  end

  def add(x)
    outdate_sum
    @array << x
  end

  def sum
    cached_sum { @array.inject {|t, e| t += e} }
  end
  
end

The code is simple. In the initialization method, there is a call to cache an attribute, sum. And then in the other methods, you'd mark where the sum would be outdated, and in the actual call to sum, you'd specify what to do to update the sum.

module AttributeCache

  def metaclass; class << self; self; end; end
  
  def cache(attr_name, options)
    instance_variable_set "@#{attr_name}", options[:initial]
    instance_variable_set "@#{:sum}_outdated", true

    metaclass.instance_eval do 
      define_method("outdate_#{attr_name}") do
        instance_variable_set "@#{attr_name}_outdated", true
      end
    end
    
    metaclass.class_eval %Q{
      def cached_#{attr_name}(&block)
        if @#{attr_name}_outdated
          @#{attr_name}_outdated = false
          @#{attr_name} = block.call
        end
        return @#{attr_name}
      end
    }

  end
  
end

The biggest hurdle was to figure out how to define a method that accepted a block with meta programming. The best I came up with was to use class_eval. If you have better suggestions, let me know. tip!