I think we should think about a general mechanism, that is for sure. That should not stop us though from using the callbacks in the meantime. OpenACS is an evolving product. No need to get it perfect right from the beginning. As long as we are aware that we should think about things and that the current implementations are bound to change, I am fine.
I am not fine with long discussions that prevent the immediate work because we say "we need to get the design absolutely right from the beginning". Code is cheap. And it is not like the callback mechanism does us much harm.
So yes, we should have a workshop and discuss the issues, but no, we should not wait for the outcome of it.
