Showing posts with label M. Show all posts
Showing posts with label M. Show all posts

Saturday, March 13, 2010

Today’s letter is M. M is for Modeling… “The Busy Developer's Guide to SQL Server Modeling”

MSDN - The Busy Developer's Guide to SQL Server Modeling

“I am a data guy. I love the nature of how data is designed, shaped and consumed. I have spent countless hours poring over database diagrams determining the right way to represent some data that I have to work with in a system. While there are a lot of different ways to do that data design, in most cases I have had to work with graphical design tools to create my data. The problem is that I am a developer. Most of the work I’ve done over the last couple of decades has been to write text that is translated into computer instructions. Working with graphical tools to perform tasks is still uncomfortable as I tend to translate the ideas I have in my head back and forth between the concept and the graphic interface. But I think in text so maybe there are other solutions that don’t require a graphical design tool.

Introducing SQL Server Modeling

Previously called Microsoft code name “Oslo”, SQL Server Modeling is a forthcoming set of technologies (presently in CTP) for solving problems around modeling data in your applications. SQL Server Modeling (or SSMod) enables you to describe your models, store them, and edit them. This represents a full life cycle for your models and the system simply becomes a runtime to consume the data contained in the models. Let’s see how this would work in a simple example.

Where we are...

SQL Server Modeling represents a new way to think about build data storage layers in your .NET projects. It introduces some different workflows that will be uncomfortable at first but in the long term being able to develop the database and data access at the same time and refactoring it along with the rest of the code should make it simpler to develop real solutions.

In addition, by utilizing the “M” language to build your models (instead of SQL), you can be more productive in building your data models. Using “M” also allows you to build both the structure (or schema) of your data model as well as reference and test data. SQL Server Modeling Services makes the design of data models first class citizens of the platform.

image …”

Oslo/M/SQL Server Modeling is something that’s I’ve been trying to wrap my head around. I’m still not sure I get it, or yet see how it’s going to help me solve my real world problems.

That said Shawn’s article HAS brought me one step closer to “getting it.” My mental fog is starting to lift…

(via Shawn Wildermuth - The Busy Developer's Guide to SQL Server Modeling)

Wednesday, December 24, 2008

Spork – Zork via “M”

Marquee de Sells - If you liked Zork, you'll love Spork!

Spork is a sample of an end-to-end application using M and the Repository. It starts by defining a set of M types that describe the data needed for a text adventure along the lines of the famous Infocom game Zork (and hence the corporate bad-café-inspired name). …”

Spork ReadMe

“Spork is a sample of an end-to-end application using M and the Repository. It starts by defining a set of M types that describe the data needed for a text adventure along the lines of the famous Infocom game Zork (and hence the corporate bad-café-inspired name). The M instances are generated by running a custom compiler developed with the VBA (Visual Basic for Adventures) MGrammar grammar. We also provide a runtime driven by adventure data loaded into the Repository in multiple versions of the types called AdvRunner.

Follow along with the video to see Spork in action or follow along with the Setup & Usage section of this document.

Requirements

You need the Microsoft Oslo SDK October 2008 CTP (aka the PDC08 drop) installed and the repository database created to run this sample. You’ll need Visual Studio 2008 installed to build it.

…”

image

Interesting…

For me M/Oslo is on a very low priority background thread, one that I’m going to let idle for a while yet, yet still I see a promise in its Textual DSL (TDSL) concept.

I’d love to be able to provide a TDSL for my business processes, allowing my dev team to focus on the actual business problems and not on the underlying goo. API’s/SDK’s/Frameworks only go so far, and, all to often, internal ones don’t go far enough yet add significant overhead, in both code and training. I’m thinking that M and TDSL’s might help with that…