> [a filter] doesn't solve any problem
Surely it solves the problem of "how do I purge my mempool of spam transactions?"
Discussion
sure, but it doesn't get rid of those transactions on the network
Like in email. Your mail server probably doesn't accept mails with an attachment of 1 Terrabyte.
Although the network and protocolls(tcp/ip, mail) support them.
There may be mailservers accepting any sized mail and even forward them, but why should you accept these obvious spam/ddos-mails?
There are other, better protocolls for big files, sane persons use ftp or torrent for things like that.
Same with the btc network.
you misunderstand something critical, you STILL get those transactions.
you get them if/when a miner mines them
but (1) you *didn't* propagate them to miners (2) you *did* reduce that miner's block propagation speed
those are two benefits you lose by opting for Bitcoin Core's mempool software
I agree with you two points above. But, unfortunately, it's is those miners with stricter filters who are disadvantaged by reduced block propagation speed 😓
That's a different problem altogether. Which you won't fix by bending over. Once Knots and better tools for decentralized block templates are the norm it will change.
"the network" is just "other people's computers." I don't care what they want to store in their mempools. I don't want to store spam in mine. *That* is the problem Knots solves.
