Grab the contents of a page by url and css selector, strip out the html, and return just the copy. When run, it checks twice, storing a copy and latest_copy. This is then used to check and see if the site has changed, only notifying when it has.

Check Please!

#!/usr/bin/env ruby
require 'rest-client'
require 'nokogiri'
require 'sanitize'

class Checker
  attr_accessor :url, :scope, :sleep_interval
  attr_reader   :page, :copy, :latest_copy
  
  def initialize(url, scope, sleep_interval)
    @url = url; @scope = scope; @sleep_interval = sleep_interval
    puts "Checking to see if #{@url} has been updated..."
    get_page
    @copy = get_copy
    check_copy
    refresh
  end

  def scope=(new_scope)
    @scope = new_scope
  end

  def get_page
    @page = Nokogiri::HTML(RestClient.get @url)
  end

  def get_copy
    Sanitize.clean(@page.css(@scope).to_html)
  end

  def check_copy
    if @copy.empty? 
      puts "There's nothing to check against. Exiting..."
      exit
    else
      puts "Here's what it says right now:"
      puts @copy
    end
  end

  def refresh
    get_page
    @latest_copy = get_copy
  end
end

site = Checker.new("http://www.rubygems.org", ".pitch", 60)
if site.copy == site.latest_copy
  puts "No change yet. We will notify you when it has."
  until site.copy != site.latest_copy do
    site.refresh
    sleep site.sleep_interval
  end
end

puts "#{site.url} posted a new update."
puts "Here's what it says:"
puts site.latest_copy
# or add your notification method of choice
# e.g. a growl notification, or a tropo.com trigger URL to send a text message, etc.