RE: [bt-spec] BTP Issue 18 : Response to CONTRADICTION ?

[Peter Furniss <peter.furniss@choreology.com>]

 

BTP Issue 18 : Response to CONTRADICTION ?

Category: minor technical
Submitter: Gordon Hamilton, AppliedTheory
Reference: Page 47 Para mid
Current text: CONTRADICTION Sent by the Superior to an Inferior that has taken an autonomous decision contrary to the decision for the atom. This is detected by the Superior when the 'wrong' one of CONFIRMED or CANCELLED is received. CONTRADICTION is also sent in response to a HAZARD message.
Question: What is the response to a CONTRADICTION supposed to be? Especially a CONTRADICTION to a hazard.

---

 

There is no response to CONTRADICTION, it's the end of the message sequence. It's receipt means the Inferior is permitted to remove the persistent information that recorded the autonomous decision - this ensures that a contradiction (the event, not the message) is known to the superior, because the Superior won't send CONTRADICTION until it has recorded (or done whatever it wants to) with the knowledge of the contradiction. (if you track through the state tables you'll see that k2 and l2, the post-contradiction states, are left only be removing the persistent information, causing a transit to state z).

 

After a HAZARD is a little messier, since there may not be any persistent information (in fact it may be precisely the inability to modify the persistent information that caused the problem). But CONTRADICTION is again the indication that the Superior has taken note of the problem, so the Inferior has no further responsibility.

 

Peter