Okay. I think am getting close to understanding where you coming from
The condition is either this or that not both which I believe rule out the fact it's merged mining.
Miners must work on separate block headers for each chain, as transactions and merkle roots differ.
Thank you, Ambatman..
The core idea is about keep a negative-hash value we may possibly find during a very normal mining process, that we regularly discard them. Just think if we keep and write them in blocks - once they get found! This is assumed there might be lots of benefits in. Just tried to explain it in more detail in reply above (to stwenhao).