AlexGR, do you allow for the possibility that maybe small blocks isn't a good idea?
Timing is crucial.
Even 1TB per year blocks (20mb/block) will have its time when 20mb/block will be "alright".
Upgrade too soon, you'll have 10gb txs and 990gb spam.
Upgrade on time, you'll get 800-950gb txs and 50-200gb spam.
I don't understand why it's spam when it pays a fee. Miners are free to exclude tx if the fee is too small.
It's not, and nobody seems to think this through. Since you can't (usually, unless through analysis or public announcement of intention) know the source or purpose of a transaction, spam in bitcoin space can only be defined as a percentage function of a transactions fee value relative to other transactions or its probability of block inclusion. So like, if a transaction is in the bottom 10% of currently average fees paid, maybe you consider that spam... or maybe you consider transactions in the bottom 50% spam. All relative to the viewer. Or maybe we are all hodlers, and none of us are actually using bitcoin for anything useful, and all of it is spam. But then you have to ask, what is useful? What if a company that is literally doing nothing but spamming the network with 'useless' transactions but is paying good fees for it and pushing useful transactions away? Is that spam?
Spam is a word that is only useful for the brain to justify something. It's not a number and bitcoin only cares about the math.