Please, be brutally honest with me. I would like to continue these presentations and there is always room for improvement.

After some lighthearted chatter outside and a filling of Meat Lover’s pizza (also known as the only pizza worthy of existence), the presentations started. Peter Mourfield kicked things off with a jQuery presentation that was pretty informative. Unfortunately, most of the crowd had a pretty good understanding of jQuery and this rendered the first portion of his presentation redundant. Peter recovered quickly and I enjoyed his “explain via slide, demonstrate in code” style.

When Peter was done it was my turn – this being my first public speaking event outside of Airman Leadership School I was a tad bit nervous, but the great crowd and atmosphere made things a lot easier (not to mention the couple cold ones I tossed back at Dave Bonds‘ house prior).

Overall, the feedback I was receiving from the crowd during the presentation and afterward let me know I was doing a pretty damn good job up there. A lot of smiles or nodding heads and very few puzzled/confused looks – if I saw those I tried to expound on that off the top of my head and you could see the question mark turn into a lightbulb over their head pretty quickly (yes, like The Sims, this is how I see people in my daily life).

I received a lot of great questions afterward and almost everyone told me they were going to check out CodeIgniter.com or ExpressionEngine.com as soon as possible.

I’ve included my slides from the presentation (via SlideShare). I fall within the “summarize on screen and explain verbally” crowd when it comes to presentations, so the meat and potatoes can be found within the slide notes (which I have conveniently placed below the presentation).

Update: Video is available!
Chris Harrison is working the video this week and I will make a separate post when that is available. Kenny Adams has already uploaded some pictures of the event as well.

Slide 2
#1 Google result for “open source PHP framework”

Slide 3
Free to download
Can virtually do anything with framework
- Can not create a commercial blogging application

Slide 4
EE2 is built on top of the CodeIgniter framework

Slide 5
EllisLab supports the company and it’s 11 employees off of ExpressionEngine primarily.
CodeIgniter is the foundation of the new version of ExpressionEngine.
EllisLab is committed to supporting the framework, it’s not a side project it is the core of their business.

Slide 6
In 2001, Rick Ellis started work on a blogging engine for his clients.
The first installation of pMachine was for Nancy Sinatra (Frank Sinatra’s daughter).

Slide 7
In 2002, pMachine was released.
Featured in Mac World and Mac Addict as well as a Blogging book published by McGraw/Hill.
Rick Ellis starts developing software full-time.

Slide 8
ExpressionEngine succeeds pMachine and is a more fully feature complete Content Management System.

ExpressionEngine sees great success within the designer market, on their personal blogs (Veerle Pieters, for instance).

Slide 9
EllisLab pulls the core code out of ExpressionEngine and refactors into a framework.

Slide 10
At SXSW ’08, EllisLab announces ExpressionEngine 2.

Derek Allard, Code Architect for EllisLab, announces that ExpressionEngine 2 will run on top of the CodeIgniter framework.

Full Circle: CodeIgniter was born from ExpressionEngine and now it is powering ExpressionEngine

Slide 11
Benchmarking performed by AvnetLabs w/ eAccelerator opcode caching

10.5x faster than CakePHP, 2.9x faster than Zend Framework (similar configurations / similar functionality)

I’ve never seen a benchmark where CodeIgniter didn’t outperform all competitors

Slide 12
Statistics dated Oct 08, from nexen.net

Shows about a 50/50 split between PHP4 and 5 support.

No doubt PHP5 has taken the lead by now, but many users still stuck on PHP4.

CodeIgniter utilizes PHP5 features/performance-enhancing functionality when present.

Slide 13
MySQL is by far the most supported but could switch without the need to completely rewrite code.
Supports MySQL, MS SQL, Oracle, PostgreSQL, SQLite and ODBC connectors.

ActiveRecord class allows you to think in PHP and generate SQL.

Slide 14
Facebook, MySpace, Twitter and other social networks have seen the rise of a new form of worm.
These worms are Javascript based, essentially the attacker enters code into a form that your site displays somewhere.
This allows the malicious code to run from your domain essentially bypassing all of the domain-lock features Javascript has built-in.

At the flick of a boolean variable, CodeIgniter will automatically protect all of your form fields from XSS vulnerabilities.

Slide 15
CodeIgniter automatically generates SEO friendly URLs

Of course, this is easily overidden: regular expression pattern matching to point to controller/method and pass variables

Slide 16
Nearly every CodeIgniter library and helper can be overridden – allowing you to increase functionality.
Examples: auto-discovery of partial views, auto-discovery of models, authentication by basing application classes off of my own controller class, which in turns extends the CodeIgniter Controller class

Slide 17
Rasmus Lerdorf at fOScon 08.

He gave a presentation pretty much slamming all PHP frameworks. If he had to pick one, it would be CodeIgniter.

Slide 18
The model represents our data – in this particular case we are returning data from a database.
Could return data from anything – a Web API, an XML file, JSON, etc.

Slide 19
All models extends CodeIgniter’s Model class and must call the parent class’ constructor from their own constructor.

Slide 20
This function starts building our SQL query by creating the ORDER BY portion.

get() will run the query against the posts table, returning 10 results starting at row 0.
We also have a get_where() method, that allows you to pass WHERE statements.
get() will return all columns, but a select() method is available to define columns if necessary.

We use the num_rows() method to return a result object or the FALSE boolean.

Slide 21
The controller accepts and handles all user input (URLs, POST/GET requests, etc.)

Like the model, a controller extends the CI Controller class.
Controllers must call their parent constructor from within their own constructor.

index() method is called by default, if a method is not defined in the URL.
We load our post model, then call the get_all() method of that model.
We then load a view, called home, and pass it our array of variables (in this case, only the posts).

Slide 22
Like the model, a controller extends the CI Controller class.
Controllers must call their parent constructor from within their own constructor.

Slide 23
index() method is called by default, if a method is not defined in the URL.
We load our post model, then call the get_all() method of that model.
We then load a view, called home, and pass it our array of variables (in this case, only the posts).

Slide 24
The view is simply an HTML file with PHP inter-mixed.
Logic should be kept to a bare minimum – just echo data in the view.
Templating engines can be used and CI even has a Template Parsing class (why would you want to add more processing time)?
Notice the lack of the $data variable – within a view, CodeIgniter brings all $data keys (or child objects) up to top-level variables.

Normally, we would check for a FALSE value before looping and show a “No Posts” message – lack of space on screen here.

Slide 25
Loop through each of the results returned from our database call (which happened in the model).

Slide 26
Echo out the title of the post, title was one of the fields within our posts table.

Slide 27
Echo out the excerpt of the post, once again a field within our table.

Slide 28
Use a CodeIgniter helper to generate a link to read the entire post.
First parameter is the URI segments, second is text of the link, third is an optional array of HTML attributes.
Slug was a field within our database – CodeIgniter has a helper to autocreate slugs from text.

Would use URI routing to process this request and send it to Controller/Method

Typically, in all of my CodeIgniter applications I implement a master view file that takes care of common code throughout all views. That master view file will then load a partial view file, based on the controller/method that is being requested. My partials always follow a specific naming convention: views/[controller]/[method].php – so it will be very easy for us to implement auto-loading of these partials.

Additionally, I follow another strict naming convention in my CodeIgniter applications, that I picked up from Ruby on Rails. Controllers are always pluralized nouns, models are named the singular version of that noun. This strict naming convention affords us the opportunity to once again auto-load a controller’s specific model.

There are a number of ways to tackle this problem, most notably hooks and hierarchical inheritance. I’ll be using the latter method in this example, since it provides a clear “breadcrumb” for others to follow as they review our application’s code. Typically, hooks are a bit more magical and don’t quite jump out at someone unfamiliar with the application and the framework it rests on.

We’re going to create our own Controller class called MY_Controller.php. This will enable CodeIgniter’s autoloading of extended libraries. We’ll then base all of our application’s classes off of MY_Controller.

We’re using CodeIgniter’s Router class to get the controller and method that is currently being called. This is a foolproof way of gathering this information, since it will remain correct regardless of any custom routing we may implement later on.

You’ll also see that we are defining a default partial of debug/echo – this will create a quick way to debug our application, in case a partial view file hasn’t been created yet.

Now we’ll create our application’s controller – I’ll be using a Users controller, since it’s something many applications implement.

We’re loading a view file named layout so let’s go ahead and create that view file (the PHP comment at the beginning is unnecessary – it’s just there to show the filename).

Finally, we’ll create the partial view file for our for our debug view and controller/method.

Now if you make a request for the Users controller you will see that our partial view is loaded correctly. If you make a request for the no_partial method of the Users controller, you should see a display of the CodeIgniter Superglobal obejct – this is because we have not created a views/users/no_partial.php view file.

Go ahead and create the appropriate model (models/user.php) and test out the autoloading of the model – it should work perfectly!

There is one downfall to this method (and it wouldn’t be that hard to fix) – you may come across some irregular plurals for your controller names (a model of mouse would have a controller of mouses). Implementing a correct reverse pluralization algorithm within the _get_model() method could solve this problem or you could just use custom routes to hide the improper grammar from your users (mapping mice to the mouses controller).

I’ve decided to just roll with the default Wordpress RSS feed, since I am not really using any of Feedburner’s features anyways. So, update your readers to use the following URL please: http://www.michaelwales.com/feed.

Definitely something I could get involved with.

Today, Chris asked if I would like to make a presentation on CodeIgniter at the May meeting. What an honor! As the CodeIgniter Community Chieftain how could I possibly turn down this opportunity to evangelize the benefits of our beloved framework.

So, what would you like to see in a CodeIgniter presentation? The reAugusta attendees aren’t new to the web development world, so I am leaning away from the “Why should I use a framework and why should it be CodeIgniter” approach – I am thinking something more along the lines of “Why CodeIgniter is Perfect for the New or Returning PHP Developer.”

Your thoughts?

Unfortunately, getting RoR up and running on Windows isn’t the simplest of tasks. Version inconsistencies between gem packages (namely, sqlite3-ruby) can turn away the most diehard nerds.

After two tries, a few uninstallations, and more than a few beers, I finally figured it out! It was pretty simple once I got all the steps down and in the right order. In the spirit of sharing (and just getting the right information out there, in case I ever need it again), the following is a step-by-step tutorial on how to install Ruby on Rails for Windows Vista. In theory, this should work on Windows XP just as well but I have not had the opportunity to test that.

The Software

We’ll be installing the following software to complete our Ruby on Rails development environment:

• Ruby 1.8.7
• Ruby on Rails 2.2
• SQLite 3.6.10

Installing Ruby 1.8.7

Installing Ruby is a pretty trivial task, thanks to the one-click installer. Unfortunately, that installer is only available for Ruby 1.8.6, and we want the latest and greatest! Nonetheless, it’s pretty simple to upgrade that 1.8.6 installation to 1.8.7.

Go ahead and download the Ruby 1.8.6 one-click installer as well as the Ruby 1.8.7 Windows Binaries. Run through the one-click installer, using the default location of C:\ruby as the installation location.

The one-click installer should add the Ruby path to your PATH environment variable – if it doesn’t you can do this within the Advanced System Settings control panel. To get to this panel, go to Control Panel > System, click Advanced System Settings, click the Environment Variables button near the bottom of the window, and prepend C:\ruby\bin; to the existing PATH variable. You can also do this through the command line: setx path “C:\ruby\bin;%PATH%”.

Now, to confirm that Ruby is installed and your environment variable is setup correctly, open the command line and type ruby --version. It should tell you that you are running version 1.8.6.

To upgrade your installation of Ruby to 1.8.7, extract the Ruby 1.8.7 Windows Binaries you downloaded and move the included files into your C:\ruby\bin directory – overwriting files of the same name. Once again, if you run ruby --version from the command line you should now be told you are running version 1.8.7.

Installing Ruby on Rails

We can use the RubyGems package manager to install Ruby on Rails. The first thing we want to do is make sure we have the latest version of the gem repository. From the command line enter gem update --system to perform an update.

Installing Ruby on Rails is just as simple. gem install rails will pull down everything you need. Finally, we’ll want to install the Mongrel server, gem install mongrel from the command line will take care of that.

Installing SQLite

The first thing we need to do is download the SQLite binaries and place them in a location that is within our PATH environment variable. We just added C:\ruby\bin to that variable, but personally I like to place utilities like this in a directory just for independent applications. I created a directory at C:\Users\[_USERNAME_]\bin (of course, replace the [_USERNAME_] portion of that with your actual username) and then I added this location to the PATH environment variable.

Download the SQLite binaries from the SQLite website and extract them into this new bin directory. I believe you only need the sqlite-3_6_10.zip file but to be perfectly honest, I just downloaded them all and placed them within my bin directory.

Finally, we need to download a gem package that provides methods for Ruby to communicate with our SQLite databases. This is the part that usually catches people when setting up RoR for Windows, as the latest version of this package does not support Windows. We’ll need to download and install an older version of the package. From the command line, enter gem install --version 1.2.3 sqlite3-ruby to download and install the latest version of sqlite3-ruby that does support Windows.

Testing

That’s all there is to it! To make sure everything worked properly, move to a clean development area within your command line (I like to use D:\rails, since my D:\ partition is my development area). Run the following commands to build a test application and start the Mongrel server for this application:
rails test
cd test
ruby script/server

After a small delay you will see that the server is now running at http://127.0.0.1:3000/ and if you visit that URL you will be presented with the Welcome to Ruby screen. Be sure to click the link for more details on your development environment and ensure no errors are displayed in this area.

You’re now ready to hop on the bandwagon and get your Ruby on Rails on!

The impact Twitter has made on my family’s upcoming move to Grovetown, GA, is immeasurable. It would have taken much more time to plan, been a lot more stressful, and I wouldn’t feel near as confident in the decisions we have made leading up to this move. Here are just a few ways I used Twitter to make the most out of our move:

1. Find Friends

Since much of the rest of this list relies on people living in the area in which you are moving to, the most important step is to find friends. Twitter makes it easy for you to search for people based on their location – remember to use city names, zip codes, and counties as well (since Twitter allows you to enter anything you like in the location field).

I’ve followed a few dozen people from the Grovetown/Augusta area and talk to 3-4 of them on a daily basis! These are people I share common interests with, enjoy the conversations we have, and would be perfectly comfortable inviting them into my home for our first face-to-face introductions.

Not knowing anyone in your new hometown can be a pretty daunting and fearful realization. By making friends, and starting a relationship, with a small group of people prior to your move you are starting out ahead of the game! If I get lost, need to know of a good restaurant, need help fixing/moving/unpacking – I already have the resources to get those issues resolved, through the network of friends I have already made. Plus, friends lead to more friends! People already in your new home’s area are sure to have other locals that they follow on Twitter – scour their timelines and watch for people they @reply to.

2. Find a Realtor

One of the most stressful aspects of moving is finding a place to live – whether you will be renting or moving. A realtor in the Augusta area is a very active Twitter user and came across my name in one of my friend’s tweets (I was actually asking people for realtor recommendations).

Although we didn’t use this realtor to find our home (we’re renting, and the home owner is represented by someone else), she proved to be a invaluable source of information about the local area. Realtors have a deep knowledge of the local neighborhoods and are the perfect place to start when you are trying to determine which neighborhood to move into. Based on your priorities (school quality, commute time to work, distance to public parks, shopping, etc.), a realtor can narrow your home search from thousands of houses down to a more manageable 5-6.

3. Selecting a House

In addition to finding a realtor, your Twitter friends can help in selecting a house. Your friends can act as a completely impartial third-party during the selection process and have no reason not to share their “gut instinct” about a property.

Ask if any of your friends live near, or drive by, the home you are looking at on a regular basis. If so, they usually won’t have a problem in scoping it out for you and making you aware of any issues they see. Plus, your friends will be able to tell you things you just can’t see within pictures (”there’s a sewage plant nearby, stinks all the time,” for instance).

4. Finding Utility Providers

You Twitter friends purchase their electricity and Internet from someone (chances are cable/satellite TV and natural gas as well). Ask them who they recommend!

Consumers are passionate about what they spend money on – especially in these trying economic times. We’re simply not going to spend money where we don’t see a valuable service. Follow your friends lead! More often than not, the majority of your friends will be using the same company and after a quick comparison and a few more questions you’ll know why (because they are cheaper, because their competitor’s service is horrible, because there is no other option).

5. Planning Your Route

The quickest route from Vandenberg AFB, CA, to Grovetown, GA, is I-40; but, that’s not the route we are taking. Ask your Twitter friends what their recommendations are and let them know your priorities (fastest route, scenic route, tourist attractions, etc.). No one knows the roads like the locals and they can warn you about congested times, construction, detours, and attractions on your route.

In our case, we wanted a quick route without any major obstacles and we wanted to stay overnight at military installations (security and affordability). We opted to drive down to I-10 (to avoid snow in the Rocky Mountains), catch I-20 in Texas and follow it all the way to our new home (Google Map of route). This meets all of our needs and only two of our seven stops will be in non-military hotels.

6. Places to Stay

Two of our stops are not on military installations so we’ll need to reserve hotel accomodations for those nights. What better way to get a recommendation than on Twitter! Ask your friends, and others that live near the city you are stopping in, what they recommend. Send them a link to a Google map showing the route you will be taking (from the previous stop to the next stop), and you are sure to get some feedback.

In our case, we wanted comfortable hotels, in the low-medium price range, close to our route, and preferably offered a continental breakfast (which opens up more “fun” money for that day’s travel). Within minutes, the La Quinta Inns in Abilene, TX, and Birmingham, AL, were recommended and after a bit of research we reserved rooms (and the price wasn’t bad either, just a few dollars above comparable military reservations).

Depending on your relationship with your Twitter friends you may even be able to pass out on someone’s couch for the night – free of charge!

I use a lot of code samples on this blog and I have found GitHub’s Gist to be an invaluable tool for doing so. It’s essentially a Pastie, where each of your Gist’s becomes its own Git repository (plus there are all kinds of other great tools as well).

Unfortunately, using Gist’s embed code generates invalid xHTML (as it doesn’t include the type parameter for the script tag). Of course, I could go through each Gist and manually add type=”text/javascript” to each of them, but what a pain in the ass.

So, I whipped up my own Gist shortcode that allows me to simply use [ gist id="1234" ] (minus the spaces inside the brackets) in my blog posts. When Wordpress finds a shortcode in a post, it sends it to a function, and outputs the returned value of that function.

To add the Gist shortcode to your own Wordpress blog, just copy-paste the following code into your theme’s functions.php (if your theme doesn’t include a functions.php just make one). Yes, this Gist is being displayed via the shortcode.

This code is pretty simple to understand. The Wordpress function add_shortcode will map the first parameter (the tag within your posts) to the second parameter (the actual function name) – in this example, they are both gist.

Within the function itself, we call extract(shortcode_atts()) which allows us to pull our shortcode’s attributes into their own local variables. Additionally, shortcode_atts() allows you to pass an array of default values as the first parameter (in case the shortcode within the post doesn’t include them). Finally, we test to make sure our $id !== NULL (meaning, the user did in fact include an ID) – if all is well, return an xHTML valid script tag which will Wordpress will then echo out, rather than our shortcode.

Now, I know those of you that read my blog via the RSS feed aren’t seeing these Gists – I’m not ignoring that! I’m working up a solution that will cURL Gist and return the raw formatted code when the RSS feed is viewed (although this may force me to leave FeedBurner and return to the standard Wordpress feed). I’m still weighing all of my options here – for now, you’ll just have to hit the domain to see the code samples.

1. The user will enter a username.
2. Query the database for the username.
3. Display the results.

Unfortunately, this is usually not the case – and is a key identifier between an entry level developer and a senior developer. Senior developers always assume the worst case scenario – a senior developer’s thought pattern is as follows:

1. The user will enter something.
2. Make sure the user’s input is the appropriate datatype.
3. Make sure the user’s input is actually a username.
   1. If so, display the user’s profile.
   2. If not, degrade gracefully and tell the user what went wrong.

Let’s take the following code, for example, which has been developed within the open source PHP framework, CodeIgniter:

/controllers/users.php

/models/user.php

/views/users/profile.php

This is a glowing example of pessimistic development. There are numerous layers of, “the user is going to mess up” within this code example.

1. This code is valid in PHP4 and 5 – we’re not assuming our user is going to make use of a particular version.
2. We’ve defined a default parameter value within our controller. If the browsing user of the site screws up, our model will return FALSE and we will show them an appropriate error message.
3. We’ve defined a default parameter value within our model. If the controller’s developer screws up, our model will return FALSE and we will show them an appropriate error message.
4. We’re checking to see if the user actually exists (if ($query->num_rows() === 1) {). If not, we will return FALSE and we will show them an appropriate error message.
5. We’re assuming the user will enter something completely crazy in our URI, but we typecast it into a string to ensure our database handles the input correctly.
6. We’re assuming the user, or the developer, will mess up by strictly limiting a successful return value to negative control statements (if ($username !== FALSE) {). By using pessimistic control statements, we ensure that the user (whether they be a developer or the browser user) must meet our exact intended parameters to receive useful data in return.

By following a pessimistic coding pattern, you will ensure that your application performs as intended each and every time – regardless of developer and browser user error. This will also force you to consider graceful degradation and useful error messages, that lead the user to achieve the results they originally intended.

As I was looking to redesign the site I really only had two requirements: 1) a wide content area, to fit plenty of code samples; and 2) a sidebar, to link throughout the domain. I found the perfect theme at ThemeForest.net and purchased it and activated it.

As always, any concerns you may have, bug reports, etc. I will definitely take into consideration. Either leave a comment here or (more effective) send me a message on Twitter, @walesmd.