[bitcoin-dev] Segregated Witness features wish list

jl2012 at xbt.hk jl2012 at xbt.hk
Sun Dec 13 15:25:10 UTC 2015

Hash: SHA256

I'm trying to list the minimal consensus rule changes needed for segwit 
softfork. The list does not cover the changes in non-consensus critical 
behaviors, such as relay of witness data.

1. OP_NOP4 is renamed as OP_SEGWIT
2. A script with OP_SEGWIT must fail if the scriptSig is not completely 
3. If OP_SEGWIT is used in the scriptPubKey, it must be the only and the 
last OP code in the scriptPubKey, or the script must fail
4. The OP_SEGWIT must be preceded by exactly one data push (the 
"serialized script") with at least one byte, or the script must fail
5. The most significant byte of serialized script is the version byte, 
an unsigned number
6. If the version byte is 0x00, the script must fail
7. If the version byte is 0x02 to 0xff, the rest of the serialized 
script is ignored and the output is spendable with any form of witness 
(even if the witness contains something invalid in the current script 
system, e.g. OP_RETURN)
8. If the version byte is 0x01,
8a. rest of the serialized script is deserialized, and be interpreted as 
the scriptPubKey.
8b. the witness is interpreted as the scriptSig.
8c. the script runs with existing rules (including P2SH)
9. If the script fails when OP_SEGWIT is interpreted as a NOP, the 
script must fail. However, this is guaranteed by the rules 2, 3, 4, 6 so 
no additional check is needed.
10. The calculation of Merkle root in the block header remains unchanged
11. The witness commitment is placed somewhere in the block, either in 
coinbase or an OP_RETURN output in a specific tx

Format of the witness commitment:
The witness commitment could be as simple as a hash tree of all witness 
in a block. However, this is bad for further development of sum tree for 
compact SPV fraud proof as we have to maintain one more tree in the 
future. Even if we are not going to implement any sum checking in first 
version of segwit, it's better to make it easier for future softforks. 
(credit: gmaxwell)
12. The block should indicate how many sum criteria there are by 
committing the number following the witness commitment
13. The witness commitment is a hash-sum tree with the number of sum 
criteria committed in rule 12
14. Each sum criterion is a fixed 8 byte signed number (Negative value 
is allowed for use like counting delta-UTXO. 8 bytes is needed for fee 
commitment. Multiple smaller criteria may share a 8 byte slot, as long 
as they do not allow negative value)
15. Nodes will ignore the sum criteria that they do not understand, as 
long as the sum is correctly calculated

Size limit:
16. We need to determine the size limit of witness
17. We also need to have an upper limit for the number of sum criteria, 
or a malicious miner may find a block with many unknown sum criteria and 
feed unlimited amount of garbage to other nodes.

All other functions I mentioned in my wish list could be softforked 
later to this system

To mitigate the risk described in rule 17, we may ask miners to vote for 
an increase (but not decrease) in the number of sum criteria. Initially 
there should be 0 sum criteria. If we would like to introduce a new 
criteria, miners will support the proposal by setting the number of sum 
criteria as 1. However, until 95% of miners support the increase, all 
values of the extra sum criteria must be 0. Therefore, even a malicious 
miner may declare any amount of sum criteria, those criteria unknown to 
the network must be 0, and no extra data is needed to construct the 
block. This voting mechanism could be a softfork over rule 12 and 13, 
and is not necessary in the initial segwit deployment.
Version: GnuPG v2


Gregory Maxwell 於 2015-12-10 04:51 寫到:
> On Thu, Dec 10, 2015 at 6:47 AM, jl2012--- via bitcoin-dev
> <bitcoin-dev at lists.linuxfoundation.org> wrote:
>> 4. Sum of fee, sigopcount, size etc as part of the witness hash tree: 
>> for
> I should have also commented on this: the block can indicate how many
> sum criteria there are; and then additional ones could be soft-forked
> in. Haven't tried implementing it yet, but there you go. :)

More information about the bitcoin-dev mailing list