Question

Last week, I downloaded source code of a blog engine developped in ASP.NET MVC3. When analyzing the code, I noticed they used a different way for retrieving data for filling the view model.

Here is an example for the ‘post’ view model:

The controller:

public ActionResult Details(string slug)
{
    var model = new PostDetailsViewModel(UnitOfWork, slug);         
    return View(model);
}

The view model:

public PostDetailsViewModel(IUnitOfWork unitOfWork, string slug)
    {
        _repository = new PostRepository(unitOfWork);
        Post = _repository.FindBySlug(slug);
    }

The repository:

    public Post FindBySlug(string slug)
    {
        return FindAll().SingleOrDefault(x => x.Slug == slug);
    }

As you can see, the repository is instantiated directly in the view model (in the constructor). Is it a right way?

Usually, I used a business layer and eventually a service layer like this:
Controller >> Business >> Repository

Thanks for your advice.

Answer 1

Is it a right way?

Doesn’t seem to me like a right way. Normally you could do this repository access in the model, not in the view model. Also they seem to have a property called Post in the view model. So they are referencing the domain model in the view model which I wouldn’t do.

As far as hardcoding a specific implementation of a repository in the constructor is concerned, well, if you don’t care about things like weak coupling between the layers, reusability, maintenance, unit testing, … you could do it. Nobody can convince us that IoC should always be done. It has its pros and cons.