Taproot Wizards launched a cartoon yesterday referred to as CatVM. I can’t discuss with it as a whitepaper, these are actual educational paperwork for adults. Within the cartoon, interspersed amongst the absurd infantile narratives, have been a number of precious technical insights relating to completely different scaling proposals within the Bitcoin ecosystem. After all, in true cartoon vogue, buried between wild exaggeration and embellishment.
The top purpose of the cartoon was to suggest a brand new mechanism for transferring out and in of scaling layers constructed on high of Bitcoin. To disentangle that precise proposal from the cartoon, we’ll have to interrupt down the 2 items concerned.
The Constructing Blocks
Rijndael’s first OP_CAT experiment was developing a vault, a scheme that enables a consumer to create an intermediate “staging” transaction to withdraw their funds from the vault. This kicks off a timelock, throughout which they will at any time ship their funds again to the vault or a safe chilly storage pockets, and after the timelock the consumer can freely withdraw the funds to the vacation spot they selected when starting the withdrawal course of. These are the solely two methods bitcoin despatched to the vault script might be spent.
Explaining the complete mechanics of how that is completed is basically an article in itself, so I’m going to do one thing I often don’t and hand waive this away as “magic.” (Defined here by Andrew Poelstra) What this “magic” permits you to do, by creating non-standard Schnorr signatures and with the assistance of OP_CAT, is to construct the transaction the signature verify is in opposition to on the script stack. This allows you to implement that sure components of the transaction are precisely as outlined forward of time. It additionally permits you to put the output from a earlier transaction on the stack within the technique of constructing the transaction spending it, which means you possibly can evaluate outputs from the spending transaction in opposition to outputs from the earlier transaction. This lets you assure by evaluating them that sure components of the earlier transaction’s outputs match sure components of the brand new outputs. I.e. the script, or an quantity. So you possibly can “carry ahead” components of the outdated outputs into the brand new ones, and implement that.
One thing else you are able to do with OP_CAT, which didn’t want Rijndael tinkering and experimenting with to show, is confirm merkle tree branches. As a result of you possibly can CAT stack objects collectively, and Bitcoin already helps hashing information on the stack, you possibly can slowly construct up a merkle tree root from a leaf node with the inside nodes. Hash two items collectively to get one hash, hash that with the pair hash, and so forth. Ultimately you get the foundation hash on the stack. You possibly can then evaluate it with OP_EQUAL in opposition to a predefined root hash within the locking script.
Unilateral Withdrawal
These two constructing blocks are sufficient to facilitate a unilateral withdrawal mechanism from a gaggle shared UTXO. A merkle root might be embedded in a transaction utilizing OP_RETURN or one other mechanism that commits to a leaf node for every consumer. The UTXO script might be structured in order that any consumer with a steadiness can try to withdraw it. To take action they would offer the merkle department committing to the quantity they’re entitled to, the authorization proof resembling a public key to verify a signature in opposition to, and assemble the transaction on the stack to confirm the suitable situations are met.
Just like Rijndael’s OP_CAT vault, this withdrawal transaction would operate as a staging level. Consumer funds could be restricted by a timelock, and they might not be able to finishing the withdrawal till it expires. At any time earlier than the timelock expires, another consumer can create a fraud proof to cease the withdrawal and shove funds again into the group UTXO script. They will do that due to OP_CAT’s capacity to confirm merkle bushes. If somebody has used a particular merkle department to withdraw funds from the UTXO earlier than, then that was included in a block someplace. By developing a transaction containing the SPV proof of that transaction inside an precise block, which might use OP_LESSTHANOREQUAL to confirm the blockheader meets some minimal issue, they will show on the stack that the merkle department was used earlier than. This enables duplicate withdrawals to be prevented.
Along with this, as a result of you need to use the “CAT on the stack” trick to make sure particular items of a earlier transaction have to be included within the subsequent, you possibly can assure that the present merkle root is carried ahead into the subsequent transaction after a profitable withdrawal. You can even assure that change from the withdrawal goes again into the group sharing script. This ensures that after one consumer withdraws their funds, the change UTXO is locked with a script that enables any remaining consumer to withdraw, and so forth. Any consumer can unilaterally withdraw their funds at any time in any order, with the assure that the rest of funds are nonetheless accessible to the remainder of the customers.
The VM Half
Readers must be accustomed to the fundamental thought of BitVM. You possibly can take an arbitrary computation and break it up into every of its constituent items and embed them in a big taproot tree, turning that computation right into a backwards and forwards problem/response recreation. This lets you lock bitcoin with extra difficult situations than is instantly supported by bitcoin script itself. The one actual shortcoming is the necessity to craft an enormous quantity of pre-signed transactions to facilitate this.
The requirement to make use of pre-signed transactions is in order that within the problem/response dynamic, you possibly can assure that cash are spent again into the big taproot tree encoding it except an exit situation in some way is reached. OP_CAT and the power to “carry ahead” information from earlier transactions permits you to assure that with no need pre-signed transactions.
So not solely does this scheme permit any consumer to unilaterally exit on their very own, it additionally permits locking situations supported by a second layer that aren’t supported by Bitcoin script to really be enforced within the withdrawal course of. I.e. if some cash have been encumbered by a sensible contract the bottom layer doesn’t perceive, after which withdrawn from the second layer, these extra difficult situations may nonetheless be settled accurately on the bottom layer because the cash are withdrawn.
The Lacking Piece
One factor that OP_CAT doesn’t allow is updating a merkle tree root representing consumer balances off-chain verifiably. It could possibly allow an already dedicated state to facilitate unilateral withdrawals, however that’s as a result of a complete part of the tree is definitely put on-chain and verified. To replace that root off-chain by definition means you aren’t placing the information on-chain. This represents an issue. There isn’t a method with simply CAT to effectively confirm that each one modifications to the merkle tree have been licensed correctly by the related customers.
Somebody(s) needs to be trusted, and by the character of issues able to spending the UTXO nonetheless and wherever they need, to effectively exchange an outdated state root with a brand new one to characterize all off-chain steadiness modifications. A brand new opcode along with OP_CAT, resembling OP_ZKVERIFY, could be wanted to do that in a trustless method.
This wouldn’t be the top of the world with out OP_ZKVERIFY although. The entity updating the merkle root for off-chain transfers could possibly be an n-of-n multisig, with 100% of the contributors required to log out on any root modifications. This boils all the way down to the identical belief mannequin as BitVM primarily based pegs, the place so long as a single sincere participant exists, nobody’s funds might be stolen. It’s a stark enchancment over present BitVM designs nonetheless with regards to the withdrawal course of.
In BitVM pegs, customers shouldn’t have a unilateral withdrawal mechanism. Peg operators have to be trusted to satisfy consumer withdrawals, figuring out that they will declare again funds they’ve spent doing so comparatively trustlessly from the BitVM peg. Whereas the incentives of this are very stable, it nonetheless does require customers primarily getting permission from another person to exit the system, they can not do it on their very own. With CatVM, customers can declare again their funds unilaterally, and an operator will not be required to entrance their very own liquidity to course of withdrawals.
Wrapping Up
Total, the design is incomplete by way of development. This isn’t one thing I’d name a Layer 2 in and of itself. It’s the core of 1, the mechanism and construction for the way funds are locked right into a Layer 2, and the method for the way customers can withdraw their funds. It undoubtedly has plenty of flexibility and usefulness to it.
Within the worst case situation, customers don’t want anybody’s permission to soundly declare their funds again on-chain. It additionally permits extra versatile programmability of funds, whereas nonetheless carrying the enforcement of these situations to the bottom layer within the occasion of worst case unilateral exits. If sooner or later we do finally get one thing like OP_ZKVERIFY, the off-chain state development can develop into an really trustless course of.
I don’t anticipate any concrete demos within the close to future, nevertheless it undoubtedly is a sound thought in my view, and one thing price contemplating. It additionally exhibits that the wizards are doing somewhat extra than simply pumping silly jpegs.