This week’s newsletter asks for comments on the miniscript language, publishes our final bech32 sending support section, includes popular Q&A from the Bitcoin StackExchange, and describes several notable changes to popular Bitcoin infrastructure projects.
- ● Evaluate miniscript: wallet developers are encouraged to evaluate this proposed language that can allow wallets to adapt to new script templates without requiring changes to the underlying wallet code for each new template. See the news section for details.
● Miniscript request for comments: the developers of this language have requested community feedback on their initial design. Miniscript allows software to automatically analyze a script, including determining what data is necessary to create a witness that fulfills the script and allows any bitcoins protected by the script to be spent. With miniscript telling the wallet what it needs to do, wallet developers don’t need to write new code when they switch from one script template to another.
For example, a wallet developer today who wants to switch from 2-of-3 multisig to 2-of-2 multisig with a 1-of-2 timelocked escape clause might have to write and test a new function for the new case. With miniscript, as long as the wallet knows how to produce signatures for specified keys and how to resolve a timelock, miniscript can guide the wallet through the various possible paths in an attempt to solve the script. No new spending code would be required.
For scripts that need signatures or other data from multiple wallets, miniscript can guide the wallet into creating all the witness data it can so that the data can be bundled into a Partially Signed Bitcoin Transaction (PSBT). Other wallets can create their own PSBTs, all of which are given to a PSBT finalizer. If the finalizer is miniscript aware, it can sort the witness data from all the provided PSBTs into a single complete witness, making the spending transaction valid.
This automation for the large range of scripts supported by miniscript allows wallets to be much more dynamic about the scripts they use, possibly even allowing users to specify their own scripts. In support of that dynamism, miniscripts can be created using an easily-written policy language. Policies are composable, allowing any valid sub-expression to be replaced by another valid sub-expression (within certain limits imposed by the Bitcoin system). For example, imagine Alice has a 2-of-3 policy that involves a hot wallet, a hardware wallet, and a fallback cold wallet:
thresh(2, pk(A_hot), pk(A_hard), pk(A_cold))
Later Alice is asked to create a fidelity bond that timelocks some of her bitcoins for a period of 26,000 blocks. The generic form of that policy is:
Alice can simply replace
pk(KEY)with her particular policy:
and(thresh(2, pk(A_hot), pk(A_hard), pk(A_cold)), older(26000))
The miniscript compiler can convert the policy into an efficient P2WSH script and check that it doesn’t violate any of Bitcoin’s consensus rules or Bitcoin Core’s transaction relay and mining policy. It can also tell Alice the various sizes related to the script so she can estimate her transaction fee expenses.
If script changes are added to the Bitcoin protocol, such as taproot, a new version of miniscript will likely be created that supports the protocol additions, making the upgrade easy for both wallets and users even if they use complex scripts.
For more information, see our miniscript topic page, C++ miniscript implementation, Rust implementation, code for a Bitcoin Core integration, and Pieter Wuille’s talk about an earlier version of miniscript (video, transcript, Optech summary).
Bech32 sending support
The last of 24 in a series about allowing the people you pay to access all of segwit’s benefits.
After six months and over 10,000 words published, this is our last bech32 sending support section. Our goal was to gently persuade as many developers as possible to add support for paying bech32 addresses in their applications. We hoped that would make it easier for segwit-ready wallets to switch from using P2SH-wrapped segwit addresses by default to more efficient native segwit bech32 addresses.
By our efforts and by the efforts of many other Bitcoiners, we think we’re on the brink of success: 19 of the 23 popular wallets and services we’ve evaluated are ready to pay bech32 addresses and 4 already generate bech32 receiving addresses by default.
Every week, it’s looking more and more reasonable for wallets to switch to bech32 soon, and we expect to hear from an increasing number of developers that their next major release will default to bech32 receiving addresses. This will lower the transaction fees for the users of that software and make more block space available to all Bitcoin users, helping to keep fees lower for everyone a little bit longer.
Even though this series has ended, we’ll continue to update the segwit section of the compatibility matrix and report on notable bech32 developments in the other parts of the weekly newsletter. We thank all of you for reading this series and for helping to improve Bitcoin scalability one wallet and service at a time.
Selected Q&A from Bitcoin StackExchange
Bitcoin StackExchange is one of the first places Optech contributors look for answers to their questions—or when we have a few spare moments of time to help curious or confused users. In this monthly feature, we highlight some of the top-voted questions and answers made since our last update.
● What are the key differences between regtest and the proposed signet? Pieter Wuille and Andrew Chow explain that while regtest is good for local automated integration tests, signet is more akin to testnet in that it allows testing of things like peer finding, propagation, and transaction selection. Signet allows for more control over block production timing than testnet and more than one signet can exist for testing different scenarios.
● Can hardware wallets actually display the amount of funds leaving your control? Andrew Chow explains that since a hardware wallet is not a full node, it needs to get its transaction amount information elsewhere. In the case of non-segwit inputs, often the amount is provided to the hardware signing device via the host computer or other wallet by sending the previous transaction to the device. In the case of segwit inputs, the amount for the input being signed must always be provided because it is a required part of the data that is signed and verified.
● How does one prove that they sent bitcoins to an unspendable wallet? JBaczuk explains that you can prove coins unspendable by sending the coins to an OP_RETURN output or another script that always returns false, or by sending coins to an address derived from a contrived, non-random script hash.
● Why is proof-of-work required in Bitcoin? Pieter Wuille explains that PoW does not create trust, but instead creates incentive for miners to cooperate with other miners by building on their blocks. PoW is also used to regulate block times (and thus protect against denial of service) since the difficulty adjustment makes it expensive to reliably produce blocks more often than every 10 minutes on average.
Notable code and documentation changes
● LND #3256 increases the number of inferences LND draws from routing failures and uses that information to adjust subsequent routes. For example, nodes are now penalized in the routing preference database if they produce an error message that they shouldn’t be generating given their particular role in a transaction (e.g. intermediate node or final node).
● BOLTs #608 provides a privacy update to BOLT4 that makes it harder for a routing node to identify which node is ultimately receiving a payment. Previously, the final node would send a
final_expiry_too_soonerror if it received a payment due to expire too soon in the future. Other non-final nodes would simply report that they didn’t recognize the payment. This made it possible to probe various channels to determine the recipient. The updated BOLT now recommends final nodes send a generic
incorrect_or_unknown_payment_detailsmessage even when they know the problem is a too-soon expiry. This is the same basic privacy leak and solution described for the wrong-amount problem in last week’s newsletter.
We thank Pieter Wuille and Sanket Kanjalkar for reviewing drafts of this newsletter and helping us understand the full range of miniscript’s capabilities. Any remaining errors are the fault of the newsletter author.