I think it was only for google auth, I personally vouch for and use keplr. Its been an amazing wallet and think the team deservers praise for it. No one is perfect, is the bottom line.
Fair enough.
The point remains, there is a tradeoff between quality and velocity. We ought to notice explicit instances where that tradeoff has gone too far in one direction. I think this is a relevant case where velocity was over-valued relative to security.
I’m not arguing that Keplr should be perfect.
I’m arguing that, if we’re discussing appropriate velocity, looking for evidence is a reasonable thing to do.
@daira is arguing that velocity is too high, hir project is (historically) zcash.
@ValarDragon is arguing that velocity is too low, what are their historical projects?
If security is the non-negotiable benchmark, then let’s at least consider relevant evidence?
Web3authn, the library that was compromised, was indeed developed against specs across multiple large tech organizations.
Specs are not a magic bullet, there is no known way to eliminate all bugs, despite much competing dogma about which way is correct.
The most reliable indicator for safe code is competent engineers who care. Process doesn’t work if you let incompetent engineers in (see Orchard bug). Minimal process works if you have the right people.
There is no way for anyone to win a debate about “proper process,” in either direction, as it’s non-falsifiable.
Personally I always bias towards working code written by proven engineers that has been subject to audits.
Having 1 person dictate a process that is both overly pedantic and unreliable is clearly not a way to move anything forward. We can’t make Zcash better like this.
Edit: @shieldedmark has said that the summary on the Arborist call was his misinterpretation/miscommunication of what you said. I apologize and withdraw my related criticisms.
I was mistaken about them actually moving to 400 zats. However, I did find the precise reason why I thought they had (see below). Also, my discussion of Zakura was entirely correct: it did implement the marginal fee as 400 zats in Zakura versions 1.7.0 through 1.8.4.
The reason why I thought Vizor had moved to 400 zat, is that in the October 1st Arborist call, @shieldedmark said the previous day he had a conversation with you from which his impression was that you planned to tell them they should. Here is a transcript, based on the Arborist call notes and checked against the video:
[Mark] … I received a message from Dev at Valar Group yesterday, signaling his intention to put the marginal fee at 400 in Zakura and start telling wallets to upgrade on October 12. Which, based on my rationale, I think I could support. If you look at the rationale in the ZIP, it’s based on the fiat price of how much it would cost to do another sandblasting attack, and since the price action has been very high —I don’t know if it’s going to sustain that height— but the rationale in the ZIP still holds. So, while his proposal is ambitious, I think it would be okay. He also seems to have influence over the Vizor wallet, which is a big wallet. So if they’re going to upgrade on October 12 and disregard the guidance in the ZIP, then we might need to scramble and get the other wallets to shift as well. […] So consider that your heads up, and we’ll do… we have inroads to other wallets, so we’ll do what we can to manage that. I understand Zodl updates on a biweekly basis, so that may conflict with that. But like I said, I’ll do what I can to liaise and hopefully get us through this rushed transition,
[Pacu] What was the date?
[Mark] They’re proposing October 12. The transaction should not fail so long as there’s a lightwalletd recipient that’s updated. At least one miner is not going to reject the transactions, and Dev’s claim is that if a sufficient percentage of the P2P network is updated, then the relay policy should still work. I dunno.
As you can see, Mark’s expectation was that Valar Group would get its way and other wallets would have to just fall into line. As it happened, in the next ZIP sync meeting on Tuesday 6th, I strongly objected to the 400 zat change being done without agreement or specification. The ZIP Editors were collectively able to persuade @ebfull that this change should be reverted in Zakura (there was another convincing technical argument against besides the objectionable politicking, but I forget what it was). That might be why Vizor never actually implemented the decrease (I had thought they had, but Claude can find no evidence of it in Vizor’s commit history), so I was mistaken in saying that “Vizor transactions were dropped by Zebra”. However, that would very likely have been what happened if I hadn’t intervened.
If Mark’s recollection of the conversation with you was inaccurate, I’m sure you can correct the record here. Otherwise, it seems a little disingenuous to say my post was “also just wrong”. Yes, Vizor didn’t actually move to 400 zats, but you had apparently planned to tell them, and possibly other wallet devs, that all wallets should do so on October 12th.
My pronouns are she/her or ze/hir (and my name is Daira-Emma). I’ll give you the benefit of the doubt that this is a simple mistake; please correct it.
Edit: since it hasn’t been corrected days later, I have to conclude it was intentional transphobia.
(Not speaking on behalf of ZF, as always)
This is the dumbest statement I have ever read in this forum. I’m deeply sad that someone that can come up with it is part of the Zcash community.
Every software has bugs and allowing one to sneak in is not a sign of incompetence. Only a handful of engineers had the knowledge to even write Orchard in the first place and most of them were already working on Zcash.
Bugs happen. I think zec had one of the finest moments in blockchain history with the way all the different dev teams came together and handled it. The disclosure was almost perfect (the zooko post on X caused some extra panic) but dev and fixing it wise, it was truly perfect. Everyone came together.
Right now we appear to have the complete opposite effect. All the dev teams and governance methods are at odds with each other. Terrible situation if there is some kind of other new problem that needs max unity again.
Would love to see everyone come together again and focus on making zec the best lifeboat we have.
It is also explicitly transphobic—misgendering Daira-Emma is not a mistake—which is the hatred and dismissal that we have been pointing out for weeks. There is a faction of self-interested men relatively insulated from real-world consequences outside Zcash that are all too eager to diminish, belittle, and overlook real contributions from trans people because they believe they, and only they, are the ones who have any real value to add to this ecosystem. It is beyond insulting and it will continue to harm this community and this project if left unchecked.
6 posts were merged into an existing topic: [ZIP] Zcash Adoption Pool (ZAP)
Tom, my AI woke me up to alert me that you are shooting from the hip and need assistance.
“Do not rush updates. Write more unit tests.”
“The testnets for future upgrades will run longer”
That is from the Osmosis June 2022 post-exploit report, after a rushed upgrade introduced a bug that allowed roughly USD 5 million to be extracted from liquidity pools.
The report called for: mutation testing, longer testnet periods and improved quality assurance.
I have no doubt Dev and his team took those lessons extremely seriously. The pressure to ship is real, especially in crypto.
Even the most highly competent engineers make mistakes under that kind of pressure.
Tom, why are you dismissing the very lessons Dev’s own project drew from a serious security failure?
Competent engineers and sound processes are not alternatives.
Sigh, at this point I think some you are just trying to push peoples buttons to get flagged.
If you wish for this particular post @thowar2 to be restored to full visibility you will need to correct the misgendering comment which is against the forums CoC.
For the record, Zakura had 400 zats at some point.
So maybe that’s where the confusion comes from?
Speaking for myself and not ZF. Disclosure: I drafted this with AI, but the views are mine.
@conradoplg agreed, and I’d go a step further. It’s close to a meme by now that Daira-Emma is always right. Of course she’s only human and she can be wrong, and she said as much herself a few posts up. But the pattern is hard to miss if you’ve been watching. She spots a problem early, she fixes it while people tell her the fix isn’t needed, and then later it turns out it was needed.
Tom H, and others arguing for faster protocol changes, please do re-read Mark’s February post on the Zcash Engineering Caucus.
“That throttling is intentional. Mainnet should change infrequently, and the bar should be high.”
Mark describes the Zcash Engineering Caucus - the core engineers who specify, build, test, review and audit the protocol. (This is a distinct group separate from the ZIP Editors):
[Zcash Engineering Caucus] are the people best positioned to tell the difference between:
- Trade-offs that are explicit, modeled, and testable
vs trade-offs that are asserted but not tied to specific failure modes- Maintenance burden that is bounded by clear interfaces + deprecation paths
vs unbounded “forever support” across clients and tooling- Attack surface defined as threat-model-first invariants, assumptions, and audits
vs “attack surface” as a rhetorical veto without specifics
And he asks:
“What does the review plan look like (ZIP review, spec review, implementation review, test vectors, audits)?”
Claims that a release is safe need evidence.
Research and development can move quickly. Mainnet activation must wait until the specification and implementations have been properly checked.
This is about protecting everyone who uses Zcash.
Mark also wrote in his February post:
“I’d love to see engineers make their reasoning legible in public”
Here are some concrete examples of why that matters.
During the NU7 release process, @str4d caught an error in how the change to 25-second blocks affected the funding-stream end heights. @nuttycom confirmed that without this review, Zcash had come close to a network consistency problem.
@daira also explained that a potential denial-of-service problem involving transaction inclusion had initially been missed.
She also described how a one-week delay to NU6.3 gave engineers time to finish formal verification work and allowed wallets and infrastructure more time to prepare.
These are specific engineering risks, not vague objections to moving quickly.
Review caught problems that could have affected everyone using Zcash. As has been said, it must come before activation, not afterwards.
This is shameful. Point me to absolutely any project and any AI will do the same “analysis” proving that any engineer is “not an expert”
have to agree with this. An AI can always nitpick and make this sort of analysis for almost every project in existence if asked to do so. +1.