Microsoft’s definition for their .NET RIA Services:

“Microsoft .NET RIA Services simplifies the traditional n-tier application pattern by bringing together the ASP.NET and Silverlight platforms. RIA Services provides a pattern to write application logic that runs on the mid-tier and controls access to data for queries, changes and custom operations. It also provides end-to-end support for common tasks such as data validation, authentication and roles by integrating with Silverlight components on the client and ASP.NET on the mid-tier.”

 I’m not too great with the technical jargon so at first sight I had no idea what it offered and failed to see what the .NET RIA Services truly offered. I began to see just how powerful it really is after reading more into it and previewing demo code.

From what I’ve gathered, the .NET RIA Services acts as the communication between server side data and the client. The communication is handled by what’s called the DomainService. It essentially provides access to whatever DataContext was provided. This reduces the need for creating any sort of Web Service on the server side and a Service Reference on the client side to consume the data. The .NET RIA Services also offer authentication services out of the box.

I’m sure there are a lot more things to get out of the .NET RIA Services, however, I’ve only recently gotten my feet wet. A couple more days reading and tinkering with demo code and I’ll be able to use this powerful tool more efficiently

The first idea that was discussed was how to build the search page. After an extensive discussion with my girlfriend, I decided that the recipes should be divided first by region e.g.  European, Asian, Southeast Asian, etc., then by cuisine, then by dish type. The reason I went with this approach is due to the way each category fits into each other.

I have no idea how many different types of cuisines there are. Doing a quick search on Google, I’ve determined that most sites will list at least 30 different types of cuisines. Then there are sites such as yelp.com that allow you to refine your search for restaurant by providing filters for 80 different types of cuisines.

Thus I saw it simpler to split the recipes into their respective regions followed by the cuisine they belong to. This will at least provide a more specific approach to the types of recipes the user will be interested in. It will also keep the attention of the user longer as a longer list would be too overbearing. I remember a rule from a multimedia class that I took a long time ago - the average user will wait about two seconds for a page to load before moving on to something else. Short and concise is key.

Filtering the cuisine by dish type is the next step in browsing for recipes. I decided to treat a cuisine as if it was your typical restaurant menu. The filters a user can choose from are:

• appetizers
• vegetables
• beef
• chicken
• pork
• seafood
• other meat
• rice
• noodles
• soup
• desserts
• drinks

Naturally, it is too difficult to categorize all the dishes into the world into just these dishes. I might possibly allow for a custom filter.

This design is far too large for a single page in the cookbook design I had earlier. I therefore decided to take this out and put it into its own page. I also decided leverage the new navigation api in Silverlight 3. The new navigation in Silverlight 3 is very powerful, robust, and easy to use. It’s definitely a great way for creating a multipage Silverlight website.

The navigation I designed is similar to the Silverlight Navigation Application template with changes to the styling. I implemented a search, profile, and login link.

The search link will bring the user to the search page. The user will be able to browse through recipes just as I have described above and add recipes to his / her liking.

The profile link will bring up the user’s cookbooks. Notice I said cookbooks. I believe adding a significant amount of recipes to a single cookbook without any means of searching will be too daunting. Therefore I decided to separate the recipes by cookbook. I’m not exactly sure how I want to go about this yet – whether the user will be able to create several cookbooks and select an active cookbook or filter the recipes into a predefined number of cookbooks based on a filter structure similar to the search.

One thing for sure is that the cookbook will take upon a cookbook design. The user can select a cookbook and flip through the pages as if it were an actual book. I’ve found some great resources and tutorials on how to implement a page flip behavior and it looks amazing.

Finally, the login link will bring you to a login page. Here I intend to leverage Microsoft’s new RIA services (still in beta by the way.) Again, I’ve found some very nice resources and tutorials on how to create a secure Silverlight website.

I haven’t written any logic for authentication just yet. I want the website to also be useable as a means for browsing through various recipes.

The implementation of the navigation system is already in place as well as the design for the login interface. The next thing I’ll be working on is the search interface and implementing a working database. Things are looking good so far.

So in my brief episodes of pondering it hit me what my project really is. This project is essentially akin to video game crafting. I’m not exactly sure how it came to mind but it dawned upon me that both contained mainly similarities.

Let’s take Aion, an MMORPG that I’ve been playing recently, as an example. A player may take upon a crafting e.g. cooking. The player then has the ability to craft a limited amount of recipes – those only within the player’s skill level. For example, the player has a cooking skill level of five. It’s possible to cook soup, which takes a skill level of one. However, it’s not possible to cook an omelet, which takes a skill level of ten. The player must “practice” cooking in order to attain a higher level of skill. More recipes can be learned as a player’s skill rises. These recipes are conveniently stored in the player’s craft’s menu. The menu contains a list of recipes that the player has learned; normally sorted by skill level. Each recipe lists the material needed to craft it as well as the amount that is created.

As you can see, the idea behind my project is similar in nature. The idea behind the user profile is essentially the same as a player’s crafting menu. The search feature of my project allows the user to learn more recipes and practice until satisfactory.

Of course the two still have their differences but I’m able to form a more concrete image of how I want this project to be presented. My project is now going to be Cookbook Builder. =)

On a completely different now. I’ve signed up with Google Code to have a means of source control. I’ve also added the VisualSVN plugin to Visual Studio 2008. This should more or less complete my development environment. Hopefully this will also be a plus as most of the companies I’ve researched used Subversion as their means of source control. All I need now is NUnit for testing and possibly continuous integration. All in good time though albeit I’m kinda limited on it.

From the outset I planned to model this project similar to a Line of Business (LOB) application. Its presentation is dependent on the data that powers it and thus I plan to incorporate Data Driven Design (DDD) and Test Driven Design (TDD.) The architecture of the project will be structured using a Model-View-ViewModel pattern (MVVM.) I also plan to leverage Ninject – a library for depency injection in .NET. Of course I’ll also be using several Silverlight libraries available to me such as the Silverlight Toolkit, Blacklight Silverlight, Silverlight Contrib + Extensions, etc.

So why all of this? Why use LOB, DDD, TDD? Why use the MVVM pattern? Why is dependency injection important? Other than to experience something different and to take on a new challenge.

I decided on functionality similar to a LOB application because of similar flow of operation. I often find my girlfriend perusing shopping sites such as UrbanOutfitter.com and AbercrombieKids.com. The workflow in such sites tend to go:

1. Search clothing
2. Add clothing to shopping cart
3. Repeat steps 1-2 as necessary
4. Sign in for checkout
5. Payment

The workflow for my project proceeds in a similar manner save for checkout and payment. A user may browse the database of recipes, select and add those of interest to the profile. From there the user can continue to browse the database repeating the same process until satisfied. From there the user may track the success of the recipe; modify the details to his/her liking and create variations; and add notes for future references. Alternatively, the user can start from the profile assuming there are already recipes added. Barring all of that, it is also possible for the user to simply use it for searching for various recipes.

The project is entirely dependent on the data that is displayed, whether it be in the form of browsing, tracking, or modify. What the users see is represented by the underlying data. Therefore I adopted the idea of DDD and TDD. I chose an MVVM pattern to aid in that thought process.

Now why MVVM? The simplest answer is coding practice.

Taking my first steps into Silverlight, I applauded its design process – separating the UI from data by placing the UI in XAML and the data in code-behind. A general rule I learned is to keep UI elements separate from the underlying logic. Having them intertwined and dependent on each other creates rigidity and a waterfall of changes. It also does not allow for isolated testing.

The answer to this is MVVM and by an extent Ninject. MVVM is similar to MVC in thought but tailored to WPF and Silverlight. The power in MVVM, just like MVC, is to divide the project into three separate components – Model – View – ViewModel. The Model is the underlying data. The View a visual representation of that data, the UI. The ViewModel is the logic; it is the bridge between the Model and the View. The Model communicates to the View via the ViewModel and vice versa.

Ninject is a third party Inversion of Control (IoC) controller. Ninject further decouples dependencies and allows for easier unit testing.

Lastly, there are an abundance of free open source Silverlight controls. Why reinvent the wheel? =)

Hopefully this is enough detail as to how I plan to approach this project. There are several great reads out there and plenty that have taught me a lot.

References:

• Jonas Follesoe’s Blog – A very good read on Silverlight, especially the use of MVVM.
• Discord&Rhyme – Nate Kohari’s, the author of Ninject, blog. General information, not to mention the uses of Ninject and IoC.
• Nikhil Kothari’s Blog – Great information in creating reusable behaviors for Silverlight.
• SilverlightShow.netTutorial on how to build a Silverlight LOB application.

The next step now that an idea of the project has been formulated is to determine the specifications of the project. What exactly do I need? Two things are absolutely necessary:

1. Personal profile
2. Food recipes

Both are obvious requirements as that would completely defeat the purpose of the project. What is next now that the minimum requirements are defined? How do we present this data?

The personal profile can be of any generic design similar to those found in online shopping web sites, forums, email, etc. The main purpose of the personal profile is to hold a record of recipes that are of interest to the user. Therefore I don’t think it necessary to display anything more than the user’s name and list of recipes.

There are two ways to represent recipes:

1. Generic recipe display. This would be seen as a way for the user to browse through the recipes and determine which is of interest. The generic recipe would contain a picture, a description, and the recipe.
2. Custom recipe display. This would be internal to the user profile. It would contain the pass or fail tag as well as notes made by the user. The generic recipe information would be included as well.

My coworker once told me to take a new project in baby steps. Focus on one thought, don’t deviate, and proceed until finished. Take a step back and make changes as necessary. Little by little. You’ll never get it perfect in one shot. You’ll only complicate it more often than not. Therefore I’ll worry about the minimum requirements for now. In the meantime, the next step is how to visually represent this data.

I believe that is due to the direction I want to take my career. I may have the most experience programming in C++ but I left the company learning C# and Silverlight. Those last moments were truly the most intriguing to me and perhaps the most challenging – learning a new language and porting an existing project to a new platform. I really did enjoy that new experience and that is what I want to pursue.

I’ll be honest. I haven’t been “actively searching” as the Department of Labor puts it. Most of my time these past 30 days were spent playing Aion Online. Here and there I would mindless wade through the many software engineering listings in the popular job pools. Other times I would stream videos on my pc. Weekends I would spend with my girlfriend and friends.

It’s been 30 days. Already 30 days since I’ve actively coded something. When you do something daily as part of a routine, it becomes part of you. It becomes embedded in your soul. It becomes so natural that often times you don’t have to think about it. It’s been 30 days since I’ve actively coded something and all those lines of code I wrote day in and day out are starting to blur. I’m losing my touch and so here we go.

I decided to blog about a personal project. This is to help me with keeping sharp, learn and increase my knowledge, and well … it also serves as a potential portfolio.

I hit a Barnes and Noble over the weekend with my girlfriend determined to start this project up. A few books I purchased to help with the learning process were

Pro C# 2008 and the .NET 3.5 Platform by Andrew Troelsen

C# Design Patterns by James W. Cooper

Pro Silverlight 3 in C# by Matthew MacDonald

Tomorrow I’ll outline the project I wish to work on and hopefully this takes off. =)