# August 7 Planning Meeting
## Attendance
Cayman
Marin
Lion
Tuyen
Patrick
Colin

## Updates
Cayman
- pushing forward gossipsub release
- added a lot of tests that mimic what golibp2p does
- some are failing, either translation error or gossipsub issue
- for monorepo, tackling various bugs and small features
- did another ssz release that fixed a bug in equals
- added ENR CLI flag options that can explicitly override different pieces of the ENR
- added Discv5 configuration setup so we selectively disable things
- we want to stop being reactive with the bugs we fix and start strategizing again, specifically, how to make things as simple as possible

Marin
- trying to fix critical bugs to give us a minimum viable clients
- core issue is sync which is broken
- V2 of regular sync is step in the right direction but our gossip validation is still a big DoS vector
- managed to fix up gossip block validation

Tuyen
- focussed on testing the sync
- found that we work with some nodes that consistently returned 0 peers
- made PR to fix this
- when we restack chain we can get a lot of peers from the beginning, looking into this further

Lion
- focussing on string of checks which will hopefully fix a bunch of issues we are experiencing
Next
- help with sync issues
- look at eth1 work

Patrick
- merged some open PR
- discussing how to make hierarchy of CLI commands more usable for the CLI docs generation we are working on
- running into issue with unit tests with CLI
- something in the lock file that made it stay there

## Planning
### Two High Level Goals:
**Improve Sync**

Cayman: leaning towards rewriting
- if we add attribution of blocks being processed to peers that sent it, it will give us a lot more option
- however, this is a pretty deep change
- if we do a rewrite, we can learn from the best aspects of other implementations
- Tuyen's idea: keep improving on current sync so we still have progress for the community
- someone else can be reworking this based on a working implementation in the background
Process: figure out best strategy - deterine what is missing from that strategy - write and alter code accordingly

**Improve Networking**
- gossip validation
- big issue is that we loose peers and we are not entirely sure why
- we have also had trouble getting peers
- peering in general needs a lot of work
- we are closer here than with sycning
- we could overhaul validation and update how some things work but we probably don't need a total rewrite

## Next Steps
Marin - continue working on gossip validation
Tuyen - continue working on current sync
Cayman - implement new libp2p update, look at sync from different implementation and see what it would take to get us there
Lion - help with sync issues - look at simplifying and optimizing eth1 work
Patrick - finish work on docks for CLI, start looking into networking

**Other work**
- we should refactor but networking and syncing should be priority
- Cayman: we could add some sort of additional metadata to add ids to jobs
- looking into grey area with block processing
- go through different clients, run locally, and see what happens when they kick us
- we can also ask in the Eth2 chat

## Long Term Ideas
- after phase 0 launch, we could look into what tweeks are possible to run Lodestar in the web
- once we have a stable phase 0 client we can start prototyping a light client
- once we build out light client, we can provide a web light client using a lot of Lodestar tooling and use it for a variety of use cases