For the next generation of i2b2, we are trying to minimize how many cells have to be locally hosted. Even more crucially, how many cells have to be locally customized and supported and how many can be generic. In other words, can we or should we reposition any abstraction barriers?
No comments:
Post a Comment