@Johnny: Agreed on getters. Very hard to do and I’ve learned to live with them. Using static analysis for preventing procedural abuse. The code is part of a small sample app I’ve been working on. Maybe at a given point I’ll put it out there somewhere.

This is the code for the ReflectionHelper class.
http://snipplr.com/view/17521/reflectionhelper/

Now I have something like this:

var security = NavigationBuilder
.ForApplication(Application.Security)
.ForTenant(Root.Tenant)
.AddPages(
PageBuilder.Named(”main”),
PageBuilder.Named(”home”).AsController(),
PageBuilder.Named(”dashboard”),
PageBuilder.Named(”administration”).ForParent(”main”).AsController()
)
.Build();

The indentation inside AddPages is simply there to make it a bit more readable.

Because i’m strugling to see any usefulness of fluent interfaces outside of places where a kind of configuration API is required. Here, done right, the fluency works very well because of it’s discoverable nature.

But to use it as a replacement for property setters – in my opinion this sounds wrong…

Any chance you could bundle the code up into a zip? It has a couple of dependencies (ReflectionHelper, BackingFieldResolver (which itself seems to depend on Mono)) and I think you have an extension method or two in there(Has and Name on Type).

Setters are the work of the devil. There’s a view that Getters aren’t too clever either but that’s a whole nother can o worms.

var doc = new Document
{
Author = “Stephen Hawking”,
Title = “The Universe in a Nutshell”,
Tags = new[]
{
new Tag { Name = “Physics” }
}
};

Where ‘Tags’ is an IEnumerable property.

Or, even better (if these things are required for the constructed object to make sense):

var doc = new Document(”Stephen Hawking”,
“The Universe in a Nutshell”,
new Tag { Name = “Physics” });

Where the third parameter is ‘params Tag[] tags’ to allow multiple tags.

Fluent methods are a syntax hack used in less expressive languages (Java, C ) and have no real place in C# 3.0 anymore.