Thanks, good to know !
So, it will be immediate when you take decision.
Will you warn and notice the community in any manner when you are ready ?

Yes.

Not completely sure, but time window is 9 milestons ~ 90 mins.

Few mins on an average PC. It was, probably, like a "long block" in Bitcoin.

We are deploying a fix that will make it easier for exchanges to process transfers, your issue will be solved semi-automatically after the fix is adopted by the spammers. You'll just need to replay the transfer after that.

Few minutes on average PC..? Then it would be much more on a low-power device like low-end ARM processors, or Atmel microcontrollers used in many devices...
If PoW cannot be done in 90 min, will TX be abandoned? If yes, that sets the min requirement for a hardware. What is the min required spec then? I am sure such calculations were done already  

A guess - probably this is resolved by network spammers doing PoW instead of device itself..?

P.S. I feel like my questions are too basic... Sorry if thats the case... Need to read that whitepaper finally

You won't find explanation in the whitepaper, it was posted somewhere in this thread, about unused area of chips which can be occupied by hashers. This is what should be implemented as ASIC - https://github.com/Come-from-Beyond/Curl/blob/master/src/cfb/curl/Curl.java - if you don't have a powerful enough processor.

No, we use a special format because it's trinary, not binary.