The W3C Credentials Community Group

Verifiable Claims and Digital Verification

Go Back


Credentials CG Telecon

Minutes for 2019-12-17

Christopher Allen: Present_
Sumita Jonak is scribing.

Topic: Introductions and Reintroductions

Introductions: none.
Reintros: Drummond from Evernym, Chief Trust Officer, chair governance wg, working on DID spec for long time (editor), involved in other standards in ccg

Topic: Announcements and Reminders

Announcements: events aw EOY, no meeting for next 2 wks
Joe Andrieu: No DID call.
Christopher Allen: DID community mtg 29JAN-1FEB in Amsterdam. 16MAR next RWOT. topics open 2mo prior (in JAN)
Joe Andrieu: Since last mtg of year, get feedback now and clear up work items. where are we now?
Christopher Allen: Number have not responded so we either archive their repo or move to JAN
Orie Steele: Thats a Github Org
Joe Andrieu: If you've worked in that repo, let's find a better home. lots of stuff in there
Orie Steele: +1 To move before archive / read only
Manu Sporny: Thought moving to CCG? not archiving...
Christopher Allen: If not an active person, then we cannot do anything
Manu Sporny: Let's be careful about archiving. did not know we were doing. http sigs ...how are we msging peo to let them know?
Manu Sporny: If ok to xfer repos, then everyone can start now. raise issue on every repo that it will be archived, give 3mo, end of march 2020
Christopher Allen: Can give till end of FEB, not 3mo
Manu Sporny: End of FEB to raise issues or will be archived

Topic: Action Items

Christopher Allen: Github to manage issues. starting at bottom...
Manu Sporny: Need volunteer to take over spec
Christopher Allen: Wo volunteer by end of FEB, we will archive. cannot add to it
Dmitri Zagidulin: What was the first repo discussed, before the JWK one?
Christopher Allen: Uport ? was volunteering ...but hasn't been worked on in 1yr. so we'll archive end-FEB
Manu Sporny: Close issue [32]
Orie Steele: This is on me
Christopher Allen: 88 -- Ongoing prob on how/where to store context for docs. where do DID methods store ...e.g. edu verifiable claims TF
Orie Steele: Wg...locs....holding off on pr ....to keep evolving ....we can close and reopen when clear path fwd.... how best to contribute to json-ld ....
Christopher Allen: Verifiable claims wg moving to....charterized....what is final approval? [poss JAN]
Drummond Reed: I already voted!
Manu Sporny: In Jan
Christopher Allen: Need process ..repos/issues need to cont in wg area. discussions/non-wg docs are in ccg. 09 JAN, end of day ET is last vote
Christopher Allen: Must get votes. ...process flows, discussions convos will happen here but it's up to members in formal wg to accept them as poss items to add to official wg
Christopher Allen: People on old list not in active wg. has that email been sent?
Dan Burnett: No. mailing list is still active, not closed. try now
Christopher Allen: Housing schema for curren json-ld sigs
Manu Sporny: Not part of json-ld grp.
Manu Sporny: Pulled into verifiable creds....will get everything for ld sigs....blv we can use security context to do today but no tutorial drafted
Orie Steele: I have a thing for doing it with github pages
Christopher Allen: Iphone app for implicit DIDs...need schema and timing for implicit DIDs
Orie Steele: Related the other issue I am working on: https://github.com/w3c-ccg/community/issues/88#issuecomment-566663424
Orie Steele: Some approaches there^
Christopher Allen: No schema for DIDS...constantly changing (btcr???) ...need to publish exts to that schema
Christopher Allen: Json schema. works with digital bazaar currently
Christopher Allen: Orie ack
Orie Steele: We have a work item for transmute + workday to use json-ld schema and context ...transmute's work w customs....ex is self-hosted in repo...
Christopher Allen: Doesnt conform to context...valid context to sign it not there
Christopher Allen: Process things ...need to revise in charter.... do in start of yr....need chairs to faciliate...need election...term is 2yrs...been in place for 3....need to update ccg processes
Dave Longley: [scribe assist by Christopher Allen]
Dave Longley: The pastebin did is correct and conforms to context....json schemas is soemthing diff....we have ld vocab
Christopher Allen: 1-Yes on context. 2- not a signed doc. ...to support it. have newer code...
Joe Andrieu: Chris said 2 things: 1- update charter; 2- elections. plan is jan.
Sumita Jonak: I am curious as to what the requirements are to formally be a part of the election process. USAA is still pending membership... hesitant to scribe because we're lurking. What are the requirements? [scribe assist by Manu Sporny]
Christopher Allen: You don't have to be a W3C Member, you just have to be a CCG member, which anyone can do. [scribe assist by Manu Sporny]
Christopher Allen: Don't have to be w3c member for ccg. can join as indiv
Christopher Allen: Can join as indiv ...lightweight community agreement

Topic: WebKMS

Christopher Allen: Existing repo under DB...should be moved since approved as work item.
Joe Andrieu: Look at DB repo, few misconfigs, don't worry, will fix.... intention is to move into repo
Manu Sporny: What we care about with any specs is that the IPR is clear. history of doc, mods/issues/etc must be preserved. ideally would transfer ownership from db to ccg. It's easy to do; could fix in header of spec. w3c ipr regime. click button then pull into ccg repo. web kms repo might cause issues. coord offline
Christopher Allen: Can delete it if needed (web kms?)
Manu Sporny: Many ways to move to preserve history.
Manu Sporny: Orie from Transmute joining as editor
Manu Sporny: May be useful to go back over spec.
Joe Andrieu: Understand how web kms intended to fit into arch
Manu Sporny: Ecosys that we're putting together has variety of components: aries...agents....digital wallets...personal data store...id hubs...enc data vaults... aspects not talked about , decentral kms. web kms. key mgmt part. clear: don't have strong feeling about web kms. it is 1 way of doing it, optional, not the only way. interoperable km.
Dave Longley: We have JavaScript open source BSD-3 licensed libraries for Web KMS here: client (https://github.com/digitalbazaar/web-kms-client/tree/http-sig) and switch, a component to plug into a server (https://github.com/digitalbazaar/web-kms-switch/tree/http-sig)
Manu Sporny: Sec 1.2 has use cases/reqs. why are we doing this spec? already km apis exist ...webauthn uses hw backed keys for dig sigs. fido keys. ...personal hw via web. web kms complementary to FIDO. look at azure, digital ocean....propreitary hw secure kms, interfaces are non-standard. so web kms tries to build UI into propreitary and open ones
Dave Longley: We still need to express the JSON schema information here: https://github.com/digitalbazaar/web-kms-switch/tree/http-sig/lib/schemas ... in the Data Model section of the Web KMS spec
Manu Sporny: Kms = black box that verfies sigs, protects keys, enc data, decrypt. simple/straightfwd
Manu Sporny: Near impossible to exfil priv keys from kms. math proves it. ex: dig sig was generated on yubikey. all matters for high-risk use cases for $$$$$. or medical lics, or academic creds, barred med lic, ....in a nutshell that's web kms. ...ways of securing tx
Christopher Allen: Same terroritory as ss keys, recovery (indiv + socioal), support multi-sig...cooperate ...quorum....multi-people/multi-sig...or just wrapper around FIDO api.
Manu Sporny: Mostly just wrapper around FIDO apis....aws, azure. foundational layer. multi-device sig...recovery....hae to care about those use cases. stuff done igher in stack, built on top of web kms
Christopher Allen: What's in your wallet, topic paper. has diagrams on pieces in id wallet. unclear where kms fits in that diagram. iterate on it and need to understand the larger wallet arch fit for web kms
Manu Sporny: Need a good diagram to show where web kms fits
Christopher Allen: The RWOT document with strawman architecture drawings: https://github.com/WebOfTrustInfo/rwot9-prague/blob/master/topics-and-advance-readings/whats-wallet.md
Jonathan Holt: Resservation: web wrappers around tools, pki, responsibility to do it right
Adrian Gropper: Separation of consent as relates to access. Curious if any ix between web kms and storage Access control protocols, or is it separate? verifiable credential visibile at level
Dave Longley: Web kms client ...switching btwn kms modules... responding to CA wrt who has key recovery control. concept of key stores in web kms w controllers to specify and how youmanage them is handled at diff layer. works cleaning w auth capabilities, sign w diff keys, perform diff ops w diff keys, delegation, http api or via web app to invoke diff kms apps
Dave Longley: It's not "browser only"
Orie Steele: Yes please! single interface for things that do key management and signing... write one integration, portability across phone, browser, azure, aws, google, etc..
Dave Longley: +1
Manu Sporny: Jonathan, correct -- huge responsiility to design correctly. heavy security eng. not web only. orie/i have been talking: poss to define api, local, used by local apps that everyone can propgram...goes to mobile, uses SE to do something. works with SE to invoke key on remote sys (another web kms) looks like local call but is to someting you control or trust. from dev persp, you don't care if api is reaching out ... it is beyond http api. interest in plugging into FIDO.
Dave Longley: WebKMS + EDVs layer and play nicely.
Manu Sporny: Not http only, local apis too. adrian, sep of concerns ...we want it betwn various subsys. expect ares agent or digital wallets will IX w web kms to do digital signing , issuing creds, accessing data hub or crypto auth
Dave Longley: WebKMS + signing VCs layers and plays nicely.
Orie Steele: WebKMS + Agents as well. its the solution to the non standard key and signature interfaces we are all using today.
Web kms designed to play nicely with all these systems. Having a standard way reduces dev burden and inc adoption rate
Dave Longley: +1 To manu. standard interface to hook up kms mod in background
Christopher Allen: Next mtg is jan 7th
Orie Steele: Thanks all :)