Well I'm glad to hear that the proposal is moving into a better
direction. Thank you for the useful feedback and suggestions.
> With the singleton approach, I suggest we continue with
> "the first element in the source must be the same
> type as the start element of the target range" and
> the rest of the usual conref validation, but we don't
> do any such validation for the target of the conrefend.
Great suggestion. After all, there's nothing for the conrefend to be
compared with in the singleton model. I agree that the recursive model
makes much more sense on the singleton model.
>
