Cykros on Nostr: My issue is less around believing Knots is particularly good at blocking spam and ...

My issue is less around believing Knots is particularly good at blocking spam and more on the configuration options. Especially around the rhetoric of "it does nothing." If it does nothing, great, leave it in, because it doesn't do anything :-).

And the argument against footguns is probably my biggest annoyance. It stinks of Apple design philosophy. If I type "rm -rf /", that is my responsibility as the owner/admin of the device, not your responsibility as the developer.

I did hear a good idea though, which was that the consensus rules could really stand to be neatly packaged into a library, to make it easier for there to be a multitude of implementations without the risk of chainsplits being as high. This, in my opinion, would be a much loftier goal of any team working on a "reference implementation" or effectively, doing open standard stewardship, rather than deciding that whatever implementation they like is the one they should ridicule anyone for not preferring themselves.

If nothing else, remembering Btcd is out there and does still have the 80 byte op_return limit was something I'm grateful for as it can hopefully help those who are in the "don't upgrade camp" find a solution between "run a node taken over by compromised interests" and "run a node by someone with eccentricities and ideas I don't align with" without being stuck with the "be left behind in case of future desireable upgrades or security fixes." Bonus, it's also what utreexo development has been primarily happening on, so it makes it easier to play with that. I don't love that datacarriersize is compiled in rather than in a config file, but, being written in Go, it's pretty trivial to just change the file and recompile.