Initiating choreographies

Hi

Just a general question to the group - does anyone have a good usecase for justifying the purpose of explicitly having to mark interactions as being able to initiate a choreography?

Due to the nature of CDL, being a multi-participant protocol description language, the first 'initiating' interaction will only be between two participants, and therefore all of the other participants can only imply that a choreography session has been initiated when they receive the first message that is intended for them.

So why do we make this special case of labelling the first interaction as being the choreography initiator - surely this can be derived from the choreography description as being the first communication(s) that occurs.

Regards
Gary