# Friday, January 30, 2009

Earlier today the Oslo SDK January CTP was released on MSDN. A lot of people have been saying since the PDC, “What is Oslo?” Oslo is a new platform from Microsoft that allows you to build data driven applications. Oslo revolves around the application’s metadata. As Chris Sells describes on in a great white paper on Oslo:

Metadata can be defined as the data that describes an application, for example, how a home page is rendered or how the purchasing workflow operates. Your application data represents the state of the execution of an application, such as what Bob put in his shopping cart on your Web site or his shipping address in your checkout workflow.

To provide a common set of tools for defining metadata so that it can be stored in a place that provides the same set of features as normal application data, Microsoft is creating "Oslo," a platform for building data-driven applications. "Oslo" is composed of three elements: a family of languages collectively called "M," a visual data manipulation tool called "Quadrant," and a data store called the repository.

Telerik is building some cool Oslo utilities and I am in the middle of designing them. As I was talking to Chris about some of the specs the other day, he asked me: “What are you using to keep track of the metadata of your application in your design process?” I was like: “Pen, paper, whiteboard, Word and Excel.” He said why are you not using Oslo? Then it struck me, I was in .NET programmer mode. So last decade. While I am using Visual Studio 2008, WPF, SQL Server 2008 and the Oslo SDK to build an application for Oslo, I was not using Oslo to help build the application.

The application is in its earliest phases (just moving from idea and drawing on a whiteboard to design.) I confess, I made my first mistake, I did not think about a model, I was thinking about the app. So I started over and started to model what the app would do using Oslo. How do you model an application using Oslo? You use the M language.

Specifically at this phase you would use the MSchema portion of the M specification.  I started by creating a schema using MSchema to hold some application artifacts. This requires a different way of thinking, but it is worth the effort because now information about my application is stored in the repository and I will have version history and a much easier time generating the application when the time comes. (You can also use the MGraph portion of the M specification to create a domain specific language (DSL), however, that part of the process won’t come for this application until a little later on.)

As I make progress designing and building this application, I will post it here. You can follow along and learn from my mistakes. Stay tuned, look for the “Oslo” category on this blog.