# CID Congress #3 - Agenda and Notes

25 June 2025
Event link: https://lu.ma/ofjr7mgd
Link to last sessions's notes: https://hackmd.io/1mXZ9HzJQWyFHNaF9yZWKg
Recording (available for 2 weeks): https://92667a81a1f139000799213716c65242.us-east-1.resend-links.com/CL0/https:%2F%2Ffiles.huddle01.media%2Frec%2Fglk-lmtl-szm%2FRM_z2jy47xTJhmp.mp4/1/01000197a80d92d3-18d91765-6f68-45ee-92cd-3963634ab641-000000/D1U0cq3jin2KO9zIq4bLWmswAM88Z4K0YFKNDFnrEXI=411

## Participants / githubs / socials :
cole anthony capilongo, mosh lee, benedict lau, jon geater, robin berjon, vmx, achingbrain, bnewbold, b5, rudy, tthomas, kate sills, bumblefudge, knowtheory, vasco

## Reference links (to check out before the call):
- https://hyphacoop.github.io/dasl-testing/
- https://dasl.ing/masl.html

## Agenda:

1. Introduction
- Content addressing for everyone.
- In CID Congress #1 (Dec) we introduced DASL.
- Meetings every 6-8 weeks with people from many different orgs and projects.
- Signal group: https://signal.group/#CjQKIA-BF01IFPTYp8-nGE-sIA9JOwNGgGRnqbdeUiP799ACEhCp9nU-VJkU1bSXB61HfYjV

1. Presentation: DASL testing suite (@makew0rld)
https://hyphacoop.github.io/dasl-testing/
https://github.com/hyphacoop/dasl-testing

1. Presentation: Rsky-satnav CAR Explorer (@rudyfraser)
https://satnav.rsky.dev/

1. Presentation: CAR indexing (@benlau & @yurko, Starling)

1. DASL Updates (@robin)
- MASL spec published https://dasl.ing/masl.html
- Integer range https://github.com/darobin/dasl.ing/issues/29
- What should we focus on in H2?

1. Data prep tooling (@jag-uk)
- Plans for improving tooling for large datasets turning into CIDs and CARs, in preparation for storage on IPFS and/or Filecoin

1. Iroh/n0des/DASL protocol (to be confirmed) (@b5)
- https://github.com/darobin/dasl.ing/issues/29

1. Intros from first-timers & other requests

1. Scheduling the next call (30 July?)

## Notes

- Cole's DASL Testing presentation
- Website: https://hyphacoop.github.io/dasl-testing/
- Github: https://github.com/hyphacoop/dasl-testing
- Already started filing issues with maintainers, that will be the focus over the next/final month of this project.
- Q&A
- bnewbold: would be curious how this would work with cborgen or other codegen; might be good to have a third state (inapplicable) beyond just pass/fail
- b5: this involved lots of manual harness creation. how can we help make this more maintainable over time?
- cole: vmx has mentioned adding this to the codec features, i could look into that
- cole: the harness is pretty simple, monorepo that also builds the website, so shouldn't be hard for people to add new libraries to it
- vmx: https://github.com/ipld/codec-fixtures @cole: we should talk about merging your work and the codec fixtures somehow. i don't think we should have two efforts doing a similar thing. (could even be as simple as codec fixtures having a script that pulls in your work automatically or so, or moving all the codec fixtures to your stuff as it has such a nice view)
- cole: That makes sense, let's talk soon

- Rudy's presentation
- Bluesky purports this credible exit path. Any user can export their data at any time, as a CAR file.
- https://satnav.rsky.dev/ lets anyone open, inspect, and diff their exported ATProto respository (hosted version is behind the local build)
- The feedback we got was folks wanted to support us, but they don't always know all the things we've built.
- Want to add lazy loading of large CARs.
- https://pol.is/report/r3nxfnmxcfmdcrusb4wdp for features

- CAR Indexing presentation (Yurko, Ben, Basile)
- Starling Lab works with USC Labs to store private files on a Filecoin node
- Needs a browser-based tool to track each file on the Filecoin network without relying on public indexing services
- Map file CIDs to their locations within fragmented CAR files (esp where metadata & data may be separated)
- Extract specific files givn either file ID or CID across multiple CAR files

- Jon's presentation on FIDL data prep tools
- https://docs.google.com/presentation/d/1C5hTei1yhXUYeCueMh3aPTSiinOMdo3Xn5Lz2mIeMgI/edit?slide=id.p#slide=id.p

- b5 on BDASL, range requests, & streaming
- p2p demo, no link

- robin on DASL
- See new(ish) MASL spec https://dasl.ing/masl.html
- https://github.com/darobin/dasl.ing/issues/29
- https://datatracker.ietf.org/doc/draft-caballero-cbor-cborc42/

#### Links shared:
- CAR determinism: https://specs.ipfs.tech/http-gateways/trustless-gateway/#car-format-parameters-and-determinism
- ATProto Sync 1.1 / inductive firehose: https://bsky.app/profile/bnewbold.net/post/3lin3bnojv22h
- WebRecorder's Custom Chunking design (not implemented): https://specs.webrecorder.net/wacz-ipfs/latest/

## 3.1

Jon: Were there any questions about my presentation?
Mosh: Jon, any plans for configuring dag width, etc. during your data prep phase?
Jon: Sort of. A lot of the project is about making conventions and not rewriting.
We can already manipulate the DAG. We don't intend to handle that in v1. We have this
on the watch-out list.

Eli: Discussion of Streamplace (stream.place)
Signed video segments using C2PA. Live stream comes in, gets slices into MP4,
embeds sign in the MP4 container, so cryptographically verified BUT still a
perfectly valid MP4.
b5: let that be the Bluesky thing!
Eli: just live now but we want to do recording. Merkle roots get interesting when
you have someone streaming 24/7
Mosh: Is the first product an atproto compatible streaming platform, or something deeply embedded in a Bluesky app like Twitter Spaces?
Eli: It's the former. We're talking about a lot of potential next steps. It's up to others to integrate.
Cole: Is the C2PA signing key related to the ATproto account key?
Eli: Post a did:key to your PDS and then we sign with that. Better integration in the future possible.
b5: What signing scheme?
Eli: Bitcoin Eth one. It's not supported by C2PA, I've forked, they keep brushing me off.
Treasury interested in interop.
Bryan: Description of why bsky picked the key system they did. Lots of hardware support and
web platform support.
Yurko: We did some early work on the adobe implementation of C2PA stream signing (not live stream this was supposed to be a different implementation) we got a PoC going with ffmpeg. Donno if its useful at all. https://github.com/eqtylab/c2pa_libffmpeg_eqty https://github.com/eqtylab/c2pa_ffmpeg/tree/c2pa

**HOT JOB ALERT**: https://jobs.stream.place/

Eli:
- Working on

Bryan:
- DASL at IETF
- signatures as part of DASL
- any goss on small go impl of DASL (eg, smaller API than IPFS go libs)

Long discussion on standards diplomacy.

Bryan: Signatures can have a few common practices that could be written up and
standardised. Best practices. Also key IDs or fingerprint of the key.
Robin: let's do it!
Bryan: great, early fall probably

Bryan: smål Go implementation of DASL would be very very nice
b5: *paging Cole*
Cole: I'm very interested in that. Seeing all those interop issues, I've been wanting to
do that.
Bryan: We might do it!
b5: I'm doing that on the Rust side, so we could have interop.
Make it production grade. Make it fast.
Bryan: Having a C implementation that could be wrapped might be fundable and useful.