This is probably a weird title, but when you face the same situation that I am facing at the moment, it will make sense. That situation is, when you are working an a web based application with a decent web tier technology, and you are trying to bind the UI layer to a back end layer. Almost all of the recent web stacks give you a declerative UI layer with some sort of markup language that allows you to refer to backing objects for bindings. In my case these objects are AOM instances (what a big surprise …).
What I am not happy with is the usual focus on backend functionality, building most of the code at the layer that is just behind the UI layer, and using the markup in UI layer just to do simple binding. Of course there is a reason that the markup in UI layer is quite simple in most recent technologies: if it was not like that you’d go back to days of ASP and JSP where code was very intimate with markup. That’s a scary mess, but I still feel that we are not making use of the bindings in the UI layer as much as we can. With some tweaking in these markup features, we may have a better distribution of functionality, with some responsibilities shifted to markup side of things. With some tooling support (cough: eclipse) this may even lead to a small domain specific language in UI layer for archetype bindings. In case of JSF, there is the Unified Expression Language extension mechanism and for WPF, there is markup extensions for XAML So in case you are using one these technologies and doing something I’m doing, you may consider taking a look at these.