Is the a reason for adding new functionality to a perfectly good app (or website), other than providing something for marketing to tweet or write press releases about? As it turns out, yes, and it hinges on the economic concept of transaction costs, in this case a user's time and monetary costs of searching for, understanding, and perhaps purchasing the new bit of behavior rather than finding it to hand (albeit perhaps less well implemented) in an already familiar setting.
In the past, I've used this argument on advocates of fine-grained object systems at the user level, and mostly been rewarded with blank stares. I put this down to my own lack of persuasiveness rather than bogosity. Now there's a nice post from John D. Cook laying out the consequences of transaction costs in an unix app setting, which I recommend as a bridge between economic and developer thinking.
(Hat tip: TJIC.)