Notes: DevCon Eth 1.x and 2.0 Meetings

Hey all, here are Dan Heyman'snotes from the meetings over the last week. I know I missed bits so please feel free to add. Also, I definitely missed people in the attendee lists. Not everyone made it onto the calendar invites. Please feel free to share as appropriate.

These are in reverse order.

11/2 Eth 2.0 Bi-Directional Needs Meeting

Attendees:

  • Danny Ryan
  • Joseph Lubin
  • Joseph Chow
  • Ben Edgington
  • Olivier Begassat
  • Stefan George
  • Joseph Chow
  • Daniel Burnett
  • Shahan Khatchadourian
  • Fredrik Harrysson
  • Peter Szilágyi
  • Rob Dawson
  • Mikhail Kalinin
  • Anton Nashatyrev

  • (?)

Minutes/Outcomes

  • Danny:
  • will plan to define the state of different spec components
  • Will get a design document in the readme to answer: what are the design goals for Eth 2.0?
  • Will get deadlines and clear request for comments from 1.0 implementers
  • Spec should come with PoCs, and will revamp efforts for this
  • Beones of Phase 0 by end of year
  • The big item is helping implementers know the state of things
  • Ben: will organize a research conference to help move research forward and encourage broader contributions

 

11/2 Eth 1.x meeting

Attendees:

  • Don’t have full list, but most people from the previous 1.x meeting
  • Piper Merriam
  • Alex Beregszaszi

Minutes:

  • Need forum/process by which we can make progress and get broad input
  • Alexey: small scale working group to get things done
  • Lane/Alexey/others: we can iterate with private→ public → private → public to benefit from small working groups making progress and still have transparent, inclusive process
  • Rough consensus that we should design a process to do this
  • Goal of process is to create EIPs + roadmap + explanatory text
  • Alexey: this shouldn’t be one big proposal, but rather a series of proposals
  • Daniel B: need a sense of urgency to get things done in a process like this
  • Agreed upon design principles
  • Small, sequential updates
  • Sense of urgency
  • As much public feedback as possible
  • Data driven analysis of problem to inform motivation for changes
  • Vitalik: Not convinced breaking changes are necessary. We should do an analysis to justify this.
  • Vitalik: uncomfortable w/ institutional private calls and absolutely against private forum
  • Lane: we can use Working Groups to move this forward?
  • Hudson: if we’re making changes that will affect DAPPs, we need a DAPP power user forum
  • Alexey: we need to decide 2 things:
  • Next steps on rent and reducing storage
  • eWasm (or other improvement to increase UX)
  • So let’s have 3 working groups
  1. Working group on rent
  1. Alexey to lead
  1. Working group on eWasm
  1. Casey to lead
  1. Working group on simulation of changes/data to inform motivation
  1. Shahan to lead
  2. Outcome is
  1. data on what happens if we don’t do anything
  2. Analysis if we’re solving the right problems
  3. Ideally, analysis of impact of solutions
  1. Work stream on reducing storage via archiving logs and blocks
  1. Peter to lead
  • We need some form of deadlines and next steps
  • June 1, 2019: target for hardfork
  • End of January event around Stanford conference
  • Joseph Chow to coordinate
  • November 30th call with members of this room
  • Hudson / Dan H to coordinate
  • Goal is for each working group to send around proposal one week in advance so everyone can do homework, give feedback
  • Each working group needs a gitter channel → email Hudson

Outcomes:

  • 3 Working Groups
  • Alexey to lead one WG1 on reducing storage
  • Casey to lead WG2 on eWASM
  • Shahan to lead WG3 on simulation
  • End of January workshop around Standford conference
  • Goal is to publish EIPs before this workshop (if not a lot sooner)
  • Joseph Chow to coordinate
  • Nov 30 call to check in
  • Proposal distributed 1 week earlier
  • Hudson / Daniel to coordinate
  • Gitter channels for each working group
  • WG leads to email Hudson

11/1 Eth 1.x Meeting

Attendees:

  • Vitalik Buterin
  • Danny Ryan
  • Joe Lubin
  • Ben Edgington
  • Nicolas Liochon
  • Hudson Jameson
  • Joseph Chow
  • Alexey Akhunov
  • Stefan George
  • Joseph Chow
  • Daniel Burnett
  • Shahan Khatchadourian
  • Jeremy Millar
  • Dror Avieli
  • Casey Detrio
  • Lane Rettig
  • Fredrik Harrysson
  • Peter Szilágyi
  • Rob Dawson
  • Mikhail Kalinin
  • Anton Nashatyrev
  • Brooke Zelenka
  • (?)

Minutes

  • 3 proposals for ETH 1.x from Peter
  1.  Can we archive blocks? Maybe store archived blocks on IPFS
  2.  Can we archive logs? Logs should be used as events but currently used as storage. Logs could now remain in the network for arbitrary number of blocks and then get archived
  3.  Is there a way to introduce rent (or other mechanism?) to bound state?
  • Peter: all clients to need to get PoW check before block processing. Fredrik agrees.
  • Fast sync has 7TB writes, a full sync is 27TB. As these grow, they will kill hardware
  • Conversation on how to implement rents:
  • Dynamic rents?
  • Fixed rents so that dapps have predictability
  • Vitalik: what gains can we get with smaller changes that don’t require the large overhaul rents would require?
  • What about reducing gas costs for non-storage endeavors? Cutting down gas refunds?
  • Alexey: also need to remember that increasing rents would incentivize people to use block space, which would slow down propagation
  • We need simulation to understand effects of changes and data to make sure we’re targeting the right problems.
  • The first 2 of Peter’s suggested changes seem like relatively easy changes. Will this cary us to 2.0.
  • Peter/Alexey: no, because they don’t bound state
  • This conversation implies that there will need to be some UX changes and some dapps might be affected by changes. Need data to prove it’s worthwhile

Outcomes:

Agreed:  reorder the operations around block processing in parity

Discuss how to move forward with specific proposals:

  a. archiving blocks

  b. archiving logs

  c. (how do we deal with the deleted information / where is it preserved)

Discuss hard and most important part: Bounded state with a proposal like rent is the most important

Action Items

  • 9am meeting to discuss 1.x process and next steps
  • 1pm meeting to discuss 2.0

11/31 Initial Meeting

Attendees:

  • Danny Ryan
  • Joe Lubin
  • Ben Edgington
  • Nicolas Liochon
  • Faisal Khan
  • Shahan Khatchadourian
  • Jeremy Millar
  • Dino Mark
  • Dror Avieli
  • Casey Detrio
  • Lane Rettig
  • Fredrik Harrysson
  • Peter Szilágyi
  • (?)

Minutes:

  • Some questions about research and how research moves to implementation
  • Current implementers want to work on parts of the spec that have been solidified but it’s hard to understand which components are ready. How to improve communications around state + reliability of spec?
  • Danny would like to get more feedback from 1.0 teams who have learned a lot from implementation. How to get more feedback and active engagement?
  • Idea: deadlines and process for feedback
  • Nicolas and Ben mention ideas of a vision document or design principles that explains the motivations for why the spec is being implemented in the way it is.
  • It’s hard to have full context on how to design and implement the spec without clarity around goals
  • There’s a difference between a spec that is implementable and whether the implementation actually achieves the goals of the specification
  • Broad set of comments to see more development in R&D. There need to be PoCs with all research in order to get implementing teams more engaged
  • There is some questions on timeline for Eth 2.0 and whether that means improvements to 1.0 should be actively happening in parallel, with a coordinated strategy

Outcomes

  • Follow-up meeting with broader audience to discuss two items:
  • Eth 1.x plans, process and strategy
  • Eth 2.0 bi-directional communication: research ←→ implementers