The common misbelieve in generative UIs
I’ve seen it over and over; the so called magic presentation services: Systems that can easily support different channels and devices (Internet, WAP, Digital-TV, …) without explicitly designing applications or meta data for them. Don’t get me wrong, I strongly believe in architectures that support multiple channels. I just don’t believe that nowadays systems can fulfill this demand in a generic or generative fashion. Since each channel or device has unique capabilities, restrictions as well as additional functionality, they have to be considered. None of the generative approaches I know take these facts into account. Most engines just render the same screens into different representation technologies and reuse the very same UI-process. The results are user unfriendly applications that can’t deal with device and channel specific characteristics.
Sometimes less is more: Support only channels and device that are important for the chosen business scenario, but the clue is to support them in a smart and efficient way.