3. My understanding is that XMR does not use a blockchain db. Working out your db structure I bleieve is why you're still developing the gui wallet. In any case what are the implications (pros/cons) re the differing db structures
This is a consequence of the unfinished Bytecoin code that we inherited, not a design decision. We are actively working on a DB (LMDB, to be accurate)
test-drive instructions.
The scenario you describe is clearly better than frequent conversions back and forth, but not as good as everything being done with shadow exclusively. In order for that to happen every merchant, exchange, etc. needs to support it directly, and you also need to be willing to except the higher inherent costs of transactions.
And then it is my understanding that you would face the same issue as Monero: no (scalable) B2B infrastructure for CryptoNote exists yet.
Would you please mind elaborating or explaining in simpler terms what a B2B infrastructure is?