I didn't play with Rapier, no. It seemed in a state of disarray, and all I really wanted to prove was that the 'guts' code worked and was a good candidate for porting to Maemo5. I believe I've nearly done that....I want to strip the CLucene dependency and then I reckon we're good to go.

I agree that the path of least resistance is an attractive route. If the Rapier python GUI looks good and functions well, let's give it a whirl. I tried the BibleTime frontend because it looked polished. It works well and has nice features.