Community Call — May 25, 2022

[expede]

Sign up: https://lu.ma/ucan

Suggested Agenda

• CACAO/SIWE
• IPLD representation
• @pdxjohnny Active Directory Gateway
• Community website

Let us know in the comments what you'd like to discuss! Questions about using UCAN, use cases in the wild, short demos, etc.

In Attendance

Please introduce yourself in the thread below if this is your first time

Video

Chat Log

01:50:05	John Andersen:	What about ed/x25519 shared key?
01:50:25	Brooklyn Zelenka (@expede):	They need things like secp256k1
01:50:33	Brooklyn Zelenka (@expede):	And BLS
01:50:51	John Andersen:	Thank you!
01:53:00	Philipp Krüger:	I think part of the problem *may* be you can't quite sign arbitrary data in wallet clients (https://eips.ethereum.org/EIPS/eip-191)
01:53:32	Benjamin Goering:	at least the pronunciation is not a MUST
01:53:48	Boris Mann:	@philipp that's a separate issue 🙂
01:53:50	Brooklyn Zelenka (@expede):	😉
01:54:01	Boris Mann:	And SIWE is attempting to standardize and get this into all wallet clients
01:54:12	Philipp Krüger:	Ok :+1:
01:55:37	Benjamin Goering:	👀 https://identity.foundation/JcsEd25519Signature2020/#jcs-ed25519-signature-2020
01:59:58	Benjamin Goering:	shoutout also to this for deterministic canonicalization/hashing https://w3c.github.io/rch-wg-charter/
02:02:37	John Andersen:	+1 to that
02:02:44	Benjamin Goering:	ucan should beat the drum on "grand tradition (from WAAY back in the SAML and WS-* days) of the Interoperability Plug-Fest”
02:02:47	Benjamin Goering:	https://lists.w3.org/Archives/Public/public-credentials/2020Apr/0198.html
02:02:57	Boris Mann:	@Benjamin -- exactly
02:03:10	Charles E. Lehner:	🤙
02:04:45	John Andersen:	Looking at doing this with GitHub with ssh keys and ed/x25519 keys
02:07:21	John Andersen:	This is the simple case with the shared key, one can extend it to do what Irakli talked about by using distinct roots to act as VCs and adding in layers of dids allowing for rotation
02:08:10	Charles E. Lehner:	Take it to Internet Identity Workshop :D
02:08:14	Benjamin Goering:	+1 CEL
02:09:15	John Andersen:	https://mermaid-js.github.io/mermaid-live-editor/edit#pako:eNpVUsFuwjAM_RUr53GBWw-TtjEG0iahAdshRVXSuG1Em1ZNsglR_n1uWgZcEr_k5fk59omltUIWsdjkrWgK2M5jY70cwC_KKf9GCdN9bABy7Qov-Zt2Sy_DSeFck1hsf7Dly-12DZsQ0x0adS80C0Kz8CwthDb8pV8v1NhkukS-oIWOPGkmgu9og6cLlgN-3l_YMJk8wkAN4Y2bnnID--tu9_negdIqUXXK56t5BIsgAoQhrY0jOxaIFRKMxCAc_MYmbNeksqcNUdBfGYtN6sDq3GiTQ-NlqVM44LED8qCzY1KhtSJH_hUgyCOUdX3oyb654ZOB-wchafcxAiWcuBoeugJadWNIrRyO-keDaEKifP2vD9bVLSqgasZm3lV1ren2S9kDq7CthFY0Lqe-jTFzBVYYs4hChZnwpYtpks5E9Q25xFelKROLMlFafGDCu3pzNCmLXOvxQpprQVNSjazzH5-i7EM
02:12:53	Benjamin Goering:	link to Gordon on evolution?
02:13:40	Boris Mann:	All of his writing here https://subconscious.substack.com/
02:13:43	Benjamin Goering:	ty
02:13:49	Boris Mann:	UCAN announced here https://subconscious.substack.com/p/layered-protocols?s=r
02:13:52	Gordon:	@Benjamin this might be what Boris was referencing https://subconscious.substack.com/p/exapt-existing-infrastructure
02:14:14	Benjamin Goering:	ty
02:15:11	Boris Mann:	https://github.com/ucan-wg/spec/pull/67
02:19:06	Boris Mann:	Biscuit for reference https://www.biscuitsec.org/
02:20:22	Brooklyn Zelenka (@expede):	Just looking at the clock, we should probably move on to ADG shortly
02:24:06	Gordon:	Analogous: HTTP header spec has a specification for the header syntax itself, but deliberately does not specify header body field.
02:24:16	Gordon:	There are separate specs for body fields.
02:24:24	Benjamin Goering:	dereference proof id to some well defined definition of ‘proof’. comparing well defined proof syntax https://medium.com/mattr-global/jwt-vs-linked-data-proofs-comparing-vc-assertion-formats-a2a4e6671d57
02:25:54	Brooklyn Zelenka (@expede):	It's a literal block chain
02:26:33	Boris Mann:	This is an emerging phrasing https://en.wikipedia.org/wiki/Inner_source
02:27:27	Boris Mann:	Flow chart from John -> https://mermaid-js.github.io/mermaid-live-editor/edit#pako:eNpVUsFuwjAM_RUr53GBWw-TtnUMpE1CA7ZDiqqkMW1Em1ZNsglR_n1uCqNcEj_n-fklzolltUIWscTkrWgK2MSJsV4O4BfllH-jhOkuMQC5doWX_E27hZchUzjXpBbbH2z5YrNZwTrEdIZG3QvNgtAslGWF0Ia_9OuVmpi9LpHPaaGUJ81U8C1t8HTFcsDPuysbJpNHGKghHLnpKSPYH3fbz_cOlFapqjMeL-MI5kEECENWG0d2LBArNLgQg3Dwm5iw3ZrKnjZEQX9pLDaZA6tzo00OjZelzuCAxw7Ig94f0wqtFTnyrwBBHqGs60NP9s2ITwbuC0LT7uMClHDiZniYCmjVXUIa5ZDqiwbRlET56l8frKtbVEC3uQzz7la3O42flD2wCttKaEXf5dSPMWGuwAoTFlGoRHtI6BudiecbsoivSlMbFu1FafGBCe_q9dFkLHKtxysp1oK-SDUkz3_lcuq1
02:30:59	Philipp Krüger:	This is what AD refers to, right? v
02:31:00	Philipp Krüger:	https://en.wikipedia.org/wiki/Active_Directory
02:31:18	Benjamin Goering:	EEE
02:31:44	Philipp Krüger:	Blogpost incoming? :P
02:32:42	Benjamin Goering:	https://docs.microsoft.com/en-us/azure/active-directory/verifiable-credentials/decentralized-identifier-overview
02:32:53	Philipp Krüger:	https://en.wikipedia.org/wiki/Active_Directory#Certificate_Services
02:33:00	John Andersen:	Just yesterday!
02:33:11	John Andersen:	Thank you!
02:33:29	Dmitri Zagidulin (@XR_Engine):	question for the group (that I'd love to ask on the next call) - has the option of /not/ using caveats been discussed previously?  
02:33:39	Sergey Ukustov:	Can we validate a capability chain, if an end consumer have no idea about an attached proof format?
02:34:04	Dmitri Zagidulin (@XR_Engine):	(the context for asking that is just - so the zCap spec started out with caveats, but  ended up not needing them (target url narrowing is sufficient) and is deprecating them)
02:34:06	Hugo @hugomrdias:	please save the chat somewhere
02:34:14	Brooklyn Zelenka (@expede):	We'll post it!
02:36:20	Dmitri Zagidulin (@XR_Engine):	I think the answer would be "no" -- it defeats the purpose of capabilities
02:38:40	Daniel Holmgren:	i have to run unfortunately. thx everyone 👋
02:38:51	Brooklyn Zelenka (@expede):	Thanks Daniel!
02:41:04	James Walker (@walkah):	Boris: most readme PRs on github 🏆
02:42:38	Benjamin Goering:	https://github.com/web3-storage/ucanto-name-system
02:42:39	Philipp Krüger:	UCANTOO
02:42:39	Gordon:	LOVE IT
02:42:47	Brooklyn Zelenka (@expede):	AMAZING
02:43:04	Benjamin Goering:	🤔 https://github.com/gobengo/activitypub-ucanto
02:44:21	Benjamin Goering:	I'll be in Austin
02:44:42	Dmitri Zagidulin (@XR_Engine):	(cheers all, have to drop)
02:44:51	John Andersen:	How do you spell Arakli?
02:44:56	Benjamin Goering:	Irakli
02:45:03	Hugo @hugomrdias:	I'm going to be in coruna for the igalia hackfest
02:45:04	John Andersen:	Ohhhh
02:45:06	John Andersen:	Duh

Notes

[ha6755ad]

I welcome the beyond spec talk idea. I have followed DIDs as well as IPLD for years. I understand the specs of UCANs by all means. What is harder to figure out exactly is which functions are best handled through the webnative SDK and tooling and when to cut a path. I am perfectly happy to use existing functionality, but struggle to draw the proper lines of what is already solved. Understanding more use cases being handled by the SDK would be most helpful.

[johnandersen777]

Would like to discuss Active Directory gateway.

Brief comment but will edit as we get closer

[cdata]

👋 Chris Joel from Subconscious, current maintainer of rs-ucan

[expede]

UPDATE: Posted the video & chat log

[johnandersen777]

https://www.phoronix.com/scan.php?page=article&item=simdjson-avx-512

Is this what you were talking about when you mentioned JSON in hardware?

[expede]

I'm trying to find it now, but there's a JSON parsing ASIC. The "simdjson updated with AVX-512" stuff you linked is obviously blazing fast and extremely well tested, but in theory someone could build a CBOR implementation with the same technique (I can't find one with quick googling, sadly). I'll keep digging for the hardware implementation though 🤔

[expede]

I don't think that this is the project, but interesting related paper where they did JSON->Arrow on an FPGA

Overall, this provides a throughput increase of up to 6x, compared to the baseline application. Also, we observe a full system energy efficiency improvement of up to 59x more JSON data parsed per joule.

...not bad