An experimental idea for pushing through a controversial bitcoin code change took another step forward this week.
The situation goes like this: with a scaling proposal favored by bitcoin’s core developers stuck in limbo (miners have so far been reluctant to signal support for SegWit), some have sought for more drastic ways to push its changes through to the network.
That’s where an upgrade process called a ‘user-activated soft fork’ (UASF) is supposed to come in.
First put forth on bitcoin’s technical mailing list a couple of weeks ago by the pseudonymous ‘shaolinfry’, the proposal outlines a new way of making upgrades to bitcoin.
The idea is complex, but the gist is that, instead of miners determining whether the change goes through, the economic players (consisting of users, exchanges, wallets and payment processors) do so instead.
It quickly sparked a discussion extending to social media about the safety of such a move. Though a new, and controversial, idea, it’s grabbed the attention of many bitcoin developers, including JoinMarket developer Chris Belcher and David Vorick, co-founder of decentralized storage platform Sia.
Then, this Sunday, shaolinfry, seemingly in a hurry to give his idea a spin, posted a second proposal.
This one puts forth a SegWit-specific UASF that fits with the requirements of the current deployment. This means that, if implemented, it could trigger the new type of soft fork as soon as October.
Shaolinfry told :
“Under that circumstance, if Segwit isn’t activated by 1st October then miners will be required to do so. If they don’t, the exchanges would reject blocks from those non-signaling miners, and the miners wouldn’t get paid.”
Belcher noted that developers are now calling it the “super-UASF”.
Backing up for a second, bitcoin needs a way to upgrade its rules if it wants to add new features like SegWit. But, upgrading is complicated for software that’s worth about $20bn, and worries of how to do it safely – like, without splitting the blockchain in two – abound.
A soft fork with support from miners is arguably the safest way to do so, since it’s backwards-compatible and not everyone needs to update.
However, shaolinfry sees a UASF as a necessary option right now.
“There has been a very dangerous myth circulating in some circles that miners can decide on the consensus rules,” shaolinfry said, adding that the bitcoin white paper describes miners’ role as limited to ordering valid transactions into blocks.
Vorick, in turn, argued that SegWit is an optional upgrade, one that no one has to use, and has benefits for the ecosystem.
He told :
“UASF is now becoming a popular idea because the miners are rejecting something that appears to have substantial community support.”
However, not convinced the method needs to be tried right away, Vorick stepped in on the mailing list to argue that a better way to implement the ‘super-UASF’ would be to first get written agreement from major economic players.
When shaolinfry was asked whether it might be too soon to roll it out, the pseudonymous programmer said:
“I personally don’t believe so, because this is not a new deployment. This BIP only needs buy-in from the exchanges, enough to hashpower into signalling, or choosing to mine something else profitable with SHA256 asics. It will take the upgraded nodes along.”
On the mailing list, shaolinfry further argued that a majority of nodes are upgraded to bitcoin software that uses SegWit. Many companies have further upgraded their software to support the change, and several Lightning Network implementations, which are dependent on it, have cropped up, offering “significant social proof” that the ecosystem supports it, the programmer said.
Besides, the developer reasoned, a UASF was previously used to deploy other upgrades without any problems.
However, many disagree that a UASF is a good idea, with discussions raising a number of questions: How does one actually measure that economic majority? Is it really safe? What happens if the miners don’t follow along?
And some think it’s just too risky right now.
PhD student Phil Daian, who works on cryptocurrency projects for Cornell’s IC3, called it the “worst of both worlds” in a detailed blog post analyzing different types of forks. His concern, which he pointed out is shared by Bitcoin Core contributor Luke Dashjr, is that, ironically, it puts too much trust in miners.
“If a single malicious block is mined, then you essentially have a chain split. And, if a majority of hash power is not enforcing the new fork, then it could be a permanent chain split where you essentially have a hard fork with the bulk hash power that’s not validating the soft fork rule. And then you have the hash power that is,” he told , adding:
“You’re relying a lot on miner honesty and good faith.”
The counterargument is that miners would have no choice but to support the chain that the economic majority wants, at least if they want to be able to continue to sell the bitcoins that they win from mining.
Vorick said more time is needed. “USAF needs more study and consideration before we can be certain. I think it’s likely that we attempt a UASF if the SegWit upgrade fails,” he said.
All that said, he’s optimistic, even if SegWit never goes through at all.
“Bitcoin never upgrading again would be a disappointing result,” Vorick said. “But I also don’t think that its spot as top dog in the cryptocurrency space would be threatened by that setback.”
“Bitcoin has captured many properties that no other cryptocurrency has.”