I just read a post from Mike Pizzo from the ADO.NET team where he announced that ADO.NET EF is pushed forward, past the orcas release and the aim is to get it as an addon for Orcas the first half of 2008. I think this is great news. Why?

When I first came across the EF I was really excited, it seemed that Microsoft finally got it. I was very really disappointed though when I tried out the February CTP. While there was a lot of talk about de-coupling the model from the storage, it seemed that the implementation of EF still was a "Bottom-Up" approach. Define your storage first, use a wizard to generate mapping and classes and then build your application.

"Surely this has to be due to the CTP status", I thought.

But first Fredrik Normen looked into it (http://fredrik.nsquared2.com/ViewPost.aspx?PostId=398) and then I got same information from Frans Bouma (http://weblogs.asp.net/fbouma/ ). The message was clear; Model first will not be supported in orcas but is targeted in the orcas + 1 time frame.

That's a bunch of rubbish. EF now looked more like a remake of Typed Datasets than anything else.

Anyways, back to the post from Mike. It seems like they have heard some of the massive feedback and the delay really looks like it's due to the fact that they now want EF to support model first scenarios.

Although there are a lot of other issues to be dealt with for EF, maybe they really are trying to do it right this time.

Check Mike's post over here: http://blogs.msdn.com/adonet/archive/2007/04/28/ado-net-entity-framework-update.aspx

I'm planning to put out a more explaining post about DLINQ, this is just a frustrated response to some of our fellow bloggers out there.

DLINQ in short is the code name for ADO.NET 3.0 which will leverage LINQ to marry objects and data.

Now, Paul Wilson and Frans Bouma already trashed the initiative in a posts here and here. My simple response is, for god sake guys!

The DLINQ presented at PDC, and in some pre-briefings for mvp's, is a technical preview of the proof of concept.

Orcas is at least 2-2 1/2 years away.

Not to be rude, but maybe you should look at the possibilities instead of clinging to your own O/R philosophies as the one and only truth out there.

 I for one love the idea that Microsoft brings O/R to all man kind.

So finally the gauge is off.

I'm going to publish a sets of blogs about a language feature that will be introduced in C# 3.0. Codename for this project is LINQ or Language Integrated Query.

In this first blog I'll basically introduce LINQ as a subject and in later posts I will drill down on different features.

LINQ is trying to address two general problem domains, the first is the idea that data usually isn't objects. Traditionally all data is saved in some kind of relational store (I know there's a lot of you out there that uses object stores but I'm ignoring you for now ;), relational data is built up of rows and columns and doesn't neccesarily materialize easily into an object.

The second general problem domain is consistency. We usually have loads of ways to write logic that selects or transfers data from one place to another. Some of the solutions doing this is really slick, others are (eh) not so slick.

So LINQ will try to solve these two (and many other) issues. Let's start of with a typical LINQ code bit;

var data = from m in typeof(string).GetMembers()

      select m.Name;

This will effectively give us a IEnumarable containing the names of strings all members. Since it's an IEnumarable it will easily get iterated over in all ways we can use it.

So what! You might say, "That's hardly revolutionary!". Well no, this is no revolution, just simplification for us as developers. Instead of having to write this code manually, we'll get som help of the compiler and the clr.

To demonstrate the power some more:

var data = from m in typeof(string).GetMembers()

      order by m.Name

      select m.Name

Well the SQL like syntax will tell you alot about what that will give us. Yes, that's right. An ordered list of member names. Now check this out as well:

var data = from m in typeof(string).GetMembers()

      order by m.Name

      where m.MemberType == MemberTypes.Field

      select m.Name

And I can go on and on. There's about 30 of theese standard operators defined in the C# 3.0 specs including billboard hits like top, skip, group by etc.the

So how does all of this fit together? Read on as I'll introduce the techniques and methods used to make this all possible. In further posts I will be talking about lambda expressions, type extensions and anonymous types.