In section 2.5.3, the syntax for mapping of roles in the bind statement implies
that the mapping is inconjunction with the variables being bound - however the
text implies that the binding of variables is independent of the binding of roles.

This needs to be clarified - e.g. if we wanted to map two variables, but only
one role, how would this be encoded?

Alternatively, there is a coupling between the variables and roles, in which
case the text is not clear.

I also think we need to have text and example to explain why binding roles is
relevant. I guess it is to make the performed choreography more re-usable - but
the only situation I can see this happening is if two separate participant
types, with their own role types, use a common interface - and then the
performed choreography references a third role type (possibly participant
independent) that also uses that same interface. In this case, both participants
can make use of the performed choreography by mapping their local role type to
the common role type. Is this the reason for having this role mapping? Is this
complexity necessary?