nomadshiba⚡⚡
· 2d
its 4MB rn.
also 1MB limit was added later.
lowering it is a softfork, don't add anything new.
I only slightly agree. Yes, the 1 MB blocksize was added later, but Satoshi added it quite literally for the purpose of stopping spam. The blocksize is logically necessary to define a fee market of finite sats competing for finite memory per Block of time. The change occurred early enough that demand for memory per Block was still below 1 MB, establishing the finite boundary before there was real demand exceeding 1MB.
Where I disagree is that the blocksize is simply 4 MB. L0 non-witness blockspace is still bounded by 1 MB. SegWit (L1) is a soft fork that redefines the block boundary by adding L1 witness data outside the L0 space, anchored back to L0, while discounting those additional bits through the 4 MWU accounting system. SegWit being a soft fork is exactly why a legacy node can remain on the same L0 chain without understanding witness, yet SegWit simultaneously attempts to redefine what a “full block” means for upgraded nodes. You cannot have it both ways: either witness is part of the Block and SegWit expanded the Block beyond 1 MB, or witness is outside the Nakamoto Block and is an additional consensus layer anchored to it.
That is why BIP141 is the problem. If you believe the Block is now effectively 4 MB and you’re actually concerned about spam and chain growth, then the first line of defense should be SegWit itself. Why are you offering discounted bits on your own node while complaining about people consuming them? A witness bit and a non-witness bit occupy identical physical memory on my SSD. Consensus can declare witness bits discounted; physics cannot. One bit does not become one quarter of a bit because BIP141 decreed it so. BIP110 does not address this problem at all. Why not limit witness and non-witness to 1MB before forking?
No one is forcing you to use the full SegWit allowance, and no one is forcing you to treat 4 MWU as your physical memory boundary. Nodes can be more restrictive in what they relay, miners can be more restrictive in the templates they build, and historical witness data can be pruned. If enough nodes refuse the discount and enough miners enforce equivalent pricing of bits in sats, that has literal consequences for the network. Finite sats once again compete against a more finite supply of memory in literal SSD space. The fee market has been weakened precisely because SegWit expanded available space while discounting part of it.
This does not solve the separate L0 consensus problem with OP_RETURN, but it makes consuming finite bitspace more expensive and reduces how much memory gets written per Block. If spam is the complaint, stop subsidizing the space being used for it and accelerate the restoration fee market to price out said behavior.
Ultimately, you are defining what proof of work means through finite satoshis competing for finite memory. Redefine the space and you redefine the measurement of work within that space. POW produces 1 Block of time/memory. Anchoring a new logical system to Genesis preserves ancestry in bits; it does not necessarily preserve invariance in logical space. Lol, I won’t even get into the fiat involved in changing the PoW algorithm and then manually adjusting protocol-defined difficulty to make the new ruler fit the old chain. Difficulty change by decree is not an honest solution to the problem.
Lowering blocksize would be a hard fork from Satoshi as your node would reject Satoshi consensus valid blocks. He defined 1 MB, my node just enforces it.
