>I guess I'm in the minority amongst computer programmers, >who seem to love producing over-arching abstractions that span multiple >things, >but I really prefer stuff that operates in a limited domain and does it >well, >to code that spans too much stuff and ends up being mediocre.
Actually, I agree with this philosophy in general. >Honestly, I think that Android native UI layer is pretty compelling - it >wouldn't be worth it to me as a developer to ignore all of the extra >stuff you get from running platform native there - things like device >rotation, integration with the home screen, pen input methods, touch >recognisation, etc, all stuff that is mobile specific, but very >necessary to deploying on a mobile platform. Agreed. It is reasonable to ask why we might want to layer another UI toolkit on top of Android. Since both Android and Pivot are Java-based, common non-UI code could theoretically be shared between a Pivot client and an Android client; though the UI code itself would be different between the two clients, it would be optimized for the capabilities of the specific deployment platform and would ostensibly provide the best user experience possible for each. What do others think?
