Reading a description of the design of the Appliance Studio's RoomWizard (now a Steelcase product sold by Polyvision) for my book, I came across this description of the tension between the capabilities of software-based devices and users needs. I think it's one of the clearest articulations of this dilemma I've read.
[O]ne of the measures of users’ enthusiasm for RoomWizard is the ease with which they generate new feature ideas. Everyone has their favourite “must have” feature.
On the one hand this is great. As suggested above, it could be taken as an indicator of the enthusiasm with which users embrace RoomWizard. It also means that we (and perhaps third-parties) will never be short of revisions, enhancements, and different versions with which to keep the product “fresh”. On the other hand, every silver lining has a cloud. In this case, the cloud has two parts: the real possibility that we might never satisfy every “need” of every customer; and the danger that we might damage our proposition--perhaps fatally--in the attempt. Of course, all these needs, and the features which purport to support them, are true and present. They’re strongly felt and very real. The question that we have to seek to answer is whether it is RoomWizard which should meet these needs, or something else. If it is RoomWizard, we need to ask the question, how?
This dilemma (that it’s easy to generate features but hard to deliver them without damaging your offering) is not unique to RoomWizard. In fact, it could probably be argued that a lot of software systems have this characteristic. It comes about because of the extent to which software can be extended and "improved" without violating any fundamental physical laws of the universe. Contrast this with a physical object, in which every new feature takes up space, introduces mechanical complexity, results in increased manufacturing cost, and so on. In the mechanical word, extra features have an obvious downside.
They then go on to discuss how they manage this tension, but I thought that this was a wise and insightful formulation, and especially prescient to ubicomp user experience design considering they wrote this more than 7 years ago.