Started writing script.

index.html

@@ -171,6 +171,8 @@ <h3>Immediately Invoked Function Expression</h3>
 
     </pre>  
 
+
+
 </section>
 
 
@@ -180,8 +182,8 @@ <h3>Immediately Invoked Function Expression</h3>
 <section class="slide" id="iife-problems">
     <h2>Problems</h2>
     <ul>
-        <li class="slide">Collisions [fender bender]<!-- are still possible --></li>    
-        <li class="slide">Coupling [legos stuck together]<!-- Can't swap to a different module --></li>    
+        <li class="slide">Namespace Collisions</li>    
+        <li class="slide">Coupling</li>    
     </ul>    
 </section>
 

script.md

@@ -0,0 +1,207 @@
+#AMD Be Lazy, Not Sloppy
+
+## What 5 things do I wan to cover?
+1. Why we need AMD (why IIFEs and namespaces are bad).
+2. How AMD works
+3. Concatenating
+4. Implementations & Adoption
+5. Plugins (CSS, text, etc.)
+
+
+## Three problems we need to solve in 2012.
+
+There are three problems that we as JavaScript need to solve in 2012. 
+
+- The first is **global pollution** [Al Gore]. That's the easy one. JavaScript
+  has a global object -- the window object, in the case of browsers -- and by 
+  default, that's where variables go. It's shared by all of the scripts on a
+  page, so we want to avoid polluting it. We don't want to put variables in
+  the global namespace because we can't be sure we're not clobbering variables
+  from another script. 
+
+  Unfortunately, we have to use some global variables so that scripts can 
+  communicate with one another.
+
+- The next problem is **our scripts are too big** [mega block]. A script
+  should have a single purpose, it should be reusable every time we need to do
+  the one specific thing that script is designed to do. It allows us to *snap
+  pieces of code together like legos*, using *design patterns* such as
+  *strategy*. As stated in the *single responsibility principle*, a script
+  should have *one reason to change*. That means a script should be so tiny
+  it defines a *single function* or a single configuration object.
+
+  Unfortunately, lots of tiny scripts are bad for the browser, because it
+  can only download up to eight scripts in parallel. We'll come back to
+  that. 
+
+  [Slide of a long list of `<script>` tags.]
+
+- We need to **resolve dependencies**. If we're going to have a bunch of tiny
+  scripts, we manually figure out what scripts have to be included on the
+  page, in what order. We need each script to make sure that any other scripts
+  it depends on is loaded first. Basically, we need an @import or #include or
+  require() statement. We already have it in Node, and it's coming in JS.Next,
+  but it's going to be a while before we can use it in the browser.
+
+
+
+## Review: What we're doing now
+
+[Slide of two functions with same-named local variable.]
+
+First, a quick review on global variables. If you declare a variable using the
+var statement inside of a function, the variable is local to that function.
+It's not visible outside of that function, and if you declare a variable with
+the same name inside a different function, there won't be any collision. The
+two can coexist happily.
+
+[Slide of a function inside of a function.]
+
+And we can put a function inside of a function. A variable in the outer 
+function is visible to inner function. 
+
+So if we want to make sure all of our variables in a file are only visible
+inside that file, and not attached to the global namespace, we can do this:
+
+[IIFE]
+
+It's just a function that wraps the whole file. It's called right after it's
+declared, and its only purpose is to contain variables so they aren't 
+polluting the global namespace. Told you that was easy.
+
+We still need to put *something* in the global namespace, so that we can 
+share information between files. But now we can be deliberate about it, by
+adding to the window object.
+
+[IIFE -- updated to add window.carousel = { ... }]
+
+That's not a perfect solution, because you can't be sure that's not another
+script on the page that defines an object named carousel, in which case one
+would clobber the other, and you have broken code, and it's not immediately
+obvious why. There's also no obvious way to fix it, because if you change
+the name of the variable in one script, you're breaking all of the other
+code that uses that script.
+
+[IIFE -- updated to add window.CLTJS.carousel = { ... }]
+
+So to get around this problem, we have made up namespaces under the host
+object. You pick a name that's unique to your company or organization or
+framework, and attach all of your objects to it. That way, your scripts
+can coexist with other scripts under a different namespace. But it's kind
+of clunky, and still doesn't solve the problem of using two different 
+versions of the same component on a page.
+
+[IIFE -- updated to put window.CLTJS.carousel in the arguments]
+
+Another good pattern is to pass the global variables in as arguments
+to the IIFE, so that a script isn't tightly coupled to the global 
+objects it depends on. So, for example, if CLTJS got renamed to MECKJS,
+or you want to use ALLY.carousel instead of CLTJS.carousel, that's only
+one line of code that needs to change.
+
+It also declares all of those dependencies up front, which is a good
+thing. You don't have to go digging through the script to figure that
+out. 
+
+Except that there's a list of local names up at the top and a list of
+global names at the bottom, and a whole lot of code  in the middle.  
+
+[Primitive version of define]
+
+Here's a simple way to fix that. Basically, take the IIFE pattern and
+formalize it into a function with a callback. (Take a few minutes
+explaining how this version of the define function works.)
+
+This is very close to the way AMD works.
+
+[More proper define with a partial implementation]
+
+This is an actual AMD module. The difference is it doesn't create 
+any global variables. Instead, it returns a value that represents 
+a module. That value gets associated with a name, which is the first
+argument of define(). 
+
+By convention, the name of the file that defines a module corresponds
+to the name of the module itself, similar to Java classes. And if
+that's the case, define() can resolve dependencies automatically. If
+a dependency hasn't already been defined, it can go grab the script
+that defines the module first.
+
+[Leave out the name]
+
+Most AMD loaders are smart enough that if you leave out the name of
+the module, it will automatically add a name based on the 
+filename. In fact, you *should* leave the name out in practice.
+
+[Leave out the depedencies]
+
+The dependencies are also optional. If a module doesn't have any 
+dependencies, you can leave that part out.
+
+[define(name?, dependencies?, definitionFunction)]
+
+Here's the signature for the define function.
+
+Now let's look at a few real world examples of AMD modules (repetition!).
+
+- return a function
+- return a string
+- return nothing
+- load jquery
+- load a specific version of jQuery
+- load a css file with !css
+- load text with !text
+- load a plain old javascript file
+- load a template (e.g. markdown)
+
+
+Dependencies are automatically resolved.
+R.js automatically resolves concatenates with dependencies.
+almond.js is an optimized, minimal loader.
+
+*Use a BMI calculator as the example. Dependencies are height and weight. BMI = scale.weight().inKilograms() / Math.pow( measuringTape.height().inMeters(), 2) 
+*Or something more trivial, like adding a rainbow border to every mention of Charlotte JS
+
+
+# Review the three problems we're trying to solve.
+
+
+# Who's using AMD?
+
+RequireJS
+Curl
+Dojo(?)
+All the libraries that use / support it.
+All the module libraries that are popping up.
+
+
+# Protips
+
+Use require() to dynamically load a dependency for polyfills, etc.
+Use plugins like css! and text! 
+Require a plain JS file with js!
+Make your own plugins.
+
+
+<h3>Me</h3>
+<ul>
+    <li>PatrickMcElhaney.com</li>
+    <li>@patrick_mc</li>
+    <li>http://github.com/pmcelhaney</li>
+    <li>http://ux.stackexchange.com/</li>
+</ul>
+
+<h3>AMD</h3>
+<ul>
+    <!-- I should put avatars here -->
+    <li>http://<strong>requirejs.org</strong>/docs/whyamd.html</li>
+    <li>James Burke - @jrburke</li> 
+    <li>Jonn Hann - @unscriptable (curl.js)</li>
+    <li>https://groups.google.com/group/<strong>amd-implement</strong></li>
+</ul>
+
+
+# Ideas 
+
+!atleast plugin (e.g. jQuery at least 1.5.1 -- if 1.7.2 is loaded, don't load another version)
+