First off, awesome book. I've been waiting for a book like this to come along for a very long time.

While I was showing the "Millett Stack" to a coworker the other day I was asked about the justification for the DTOs and not letting the model bleed up to the UI. I couldn't remember why you did this so I went back to your book to look it up and couldn't find the justification. I think perhaps it's just so the UI layer is only dependent on the Service layer? Just kind of want to hear the formal explanation from the man himself. If you'd prefer to reference the book I'm fine with that as well. BTW, We're actually using EF code first (clean POCO) with EF layer separated in the repository layer within the stack.

Thanks for again for a great book!

Hi ludwigs3rd,

Thanks for reading the book!

I cover ViewModels on page 349. Sometimes the UI needs to show a summary of different parts of the domain model. Passing back a full model for the UI would be too clunky, better to have a ViewModel specific to the view. If however you have a nice one to one mapping between the UI and the entity then you can just return the model. Does that answer your question?

I prefer now to follow the CQRS pattern and have a query service separate from my domain model to provide the ViewModels/Dtos for my views. This enables my domain model to only be concerned with enforcing business rules and not how it's displayed. This is something I will cover in my new book. I am on iPad at mo but will knock up a quick diagram when I'm home to show you what I mean.

Let me know if you want any more info or if you have any questions.

Cheers
Scott