I think replace by fee is good. It is a way to unstick a transaction that is stuck with too low fee, when you are in a hurry.  It does not change the protocol in any way, it is just the miner chooses the transaction with the higher fee, making the old one illegal. It does not fill up blocks, only the network.

On the other hand, I don't think it is critically important, necessary or even specially useful. I think the market will sort out fees, making normal high-paying transactions fill up only half the space available in blocks.

They cannot handle 11.6MB unconfirmed transactions then how will they handle 20 MB blocks with 200 MB uncorfirmed transactions ?

It would be dramatically more expensive to spam the network with 200 MB of transactions.

plus, miners have the power to choose to process the spam or not.  process it, make lotsa fee income thus strengthening themselves (which is the last thing a spammer wants), but also risk getting orphaned with given today's connectivity rates.  or choose not to, protecting themselves from an orphan, and continue processing small efficient blocks for maximum propagation. 

but also way more expensive for Bitcoin's growth prospects as new users get turned away or fed up.  Cripplecoin, in the long run, fails to the benefit of Blockstreams SC project.

you keep ignoring the fact that miners can filter spam; if they want.

no,look at the current spam attacks.  

they are small tx's with the minimum fee.  if that spam wants to actually get included in the next block or 2, it has to increase the amount of fees it pays which would cause the attack budget to skyrocket.

this is precisely why we disagree on everything.  i believe that the internet can allow the ppl to take back control of their money as it should be.  don't forget that central banks are still a relatively new concept in the modern age.  with that, i mean, going back 100 yrs or so. 

maybe it's naive, maybe not.  but it certainly is worth the effort and risk to try.  which is why i support lifting the block limit just to see what happens.  i believe we will cope.

we are consistently above the 2.3 TPS level that the Tradeblock ppl have defined as the level where the network experiences delays.

from Tradeblock:

(2): It appears transaction confirmation will begin to experience notable delays at roughly 2.3 TPS – the rate at which average transaction processing time begins to exceed 1 block.

To be clear, I agree. Smoothly adjusting fees are another tool in the scaling toolbox, though, and should be ready for use as well. Where we're going, a 8 or 20x increase all by itself might not cut it. I'm eyeing a 500x price increase for the next bubble (peak, crashing afterward to 5-10x lower: ~$10-20K/BTC), which will mean a pretty big increase in transaction volume and may overrun even the 20MB blocks if we don't even have a fee market up yet.

We need as many scaling solutions in place as possible to enable the next bull run.

The nice thing about this debate is it's incentivizing them all. I'm confident we'll have the blocksize cap increase, hopefully in advance and not as a reaction to a stalled rally, but also a lot of other things. It's worth pushing against the small-blockers not because they have any real power to stop the increase, but because it lights a fire under them to do their part in making those optimizations to try to prove we can get by with smaller blocks. It's a futile effort on their part, but it ends up helping with large blocks, too.