Daniel Fagerstrom wrote:
components and property DI for optional components might be a good approach. However, IMHO, it would not give enough advantages to be worth any back compatibillity, a prolonged period of instabillity or (at least for me), the effort.
Thanks Daniel for speaking it out loudly: Who wants to get his hands dirty with the two implementations
- kernel (inter-block communication, classloader shielding) - new Cocoon core engine
and which time frames are those people thinking of?
If we are only talking "theoretically" it is a very boring discussion ...
-- Reinhard
