So the definition of spam is something that is bought cheap?
The point is that it doesn't matter if it's free or paid, if what you pay is near-zero cost. It definitely cannot be a factor of ruling out spam if the fees are too cheap and aren't an adequate deterrent.
And even "adequate deterrent" is wrong, as a term, because a script kiddie might not afford a good spam attack but a deep-pocketed adversary may not be deterred by the costs, because by attacking in this fashion he is getting side-benefits by harming BTC.
From these graphs I conclude that about 2/3rd of all tx are part of chains longer than 10. So what?
When you receive a payment, do you often send your coins to another address, to another address, to another address, to another address, in a chain of hundreds or thousands? Are we serious?
Do you want people to just sit and pretend that all this is somehow "natural growth" and that there is a "problem" when the majority of txs are junk - done by god knows who - and we should give him/them the tools to continue doing that at ever lower prices, and with more space to spam?
If you have an attack vector, or a weakness, do you amplify it?
Is this rational behavior?
But anyway, this issue is ..."solved" because either classic or core, 1mb is dead and we go to 1.7 to 2.
Question is, what should the minimum fee be at a given block size limit to have enough headroom for a consistent user experience? Recycling an arbitrary limit as a capacity quota is not a good answer.
Things like that aren't simple. Someone could counter-argue that the central planning of fees isn't much different than central banking policies.