> The question of statement cancellation has been discussed at length
> with a fairly broad acceptance of the token solution.

To avoid wasting anyone's time with already-answered questions, can someone give me a pointer to a previous thread that identifies how the "token" mechanism is supposed to work, based on this broad acceptance? A generalization that identifies "what" is to be cancelled would not adversely affect our needs at all, and I do expect that the remote case (which I don't pretend to understand) will result in considerable change to the proposal.

Setting aside the mechanism for identifying the target of a cancellation operation, I am confused as to why a statement cancel isn't equally useful for a remote scenario as it is for the embedded case. Are all remote applications really essentially a 1:1 binding of attachments to statements? I.e. is there no practical difference for a remote application in cancelling a long-running statement and aborting the attachment and everything that goes with it?

I guess I'd advocate that whatever ultimately is done, it really should include the ability to cancel a specific statement.

- Tom