just committed some initial receive and correlation ideas. not complete yet. flow test still runs ok.
now i'm going to try and find out about how to parse/store the correlations in the model.
i already know that the correlations referenced in the receive fit very well onto the descriptor persistence model that we have. which means that a List correlations will probably be part of the receive executable.
The correlation objects are then objects without a reference to any other bpel process element. They only refer to the correlation sets by name, just as in the spec.
i'll be digging some more in that later after my meeting. just wanted to keep everyone posted.