JR

>Bonnie, what do you think of this: http://udidahan.weblogs.us/2007/04/21/domain-model-pattern/

Sorry, I ain't Bonnie :)

>The use of DataSets in .NET is often a sign of the Anemic Domain Model Anti-Pattern.

I found this quite interesting, relative to his definition of the "Anemic Domain Model Anti-Pattern". If I am understanding what he is trying to say, whilst I could kind of see where he is coming from with pre .NET 2 datasets, with the advent of partial classes in .NET 2, it is relatively trivial to create a partial class belonging to the dataset that contains any/all manner of logic and behaviour related to the data objects in the dataset. This also provides extended advantages that were not available prior to .NET 2. For example, if/when the dataset is regenerated, you don't lose your additional logic because it is retained in your own partial class file.

So, I am not totally sure that the reference actually holds good in this instance. On saying that, I am a fan of the domain model and helped build a .NET application development framework based on the domain model that took many leads from Martin Fowler's ideas. However, I think datasets are quite valid and often misunderstood :)

Cheers