The W3C Credentials Community Group

Meeting Transcriptions and Audio Recordings (2014-today)

Go Back

Verifiable Traceability Task Force

Transcript for 2024-02-20

Our Robot Overlords are scribing.
Russell_H_( Yeah I can scrape.
Mahmoud Alkhraishi: Awesome thank you Russell.
russell_h_( is scribing.
Mahmoud Alkhraishi: Welcome
Patrick St-Louis: Can close this, another pr fixed this issue. Open a new one to decide what to do with empty contexts, but this is resolved.
Mahmoud Alkhraishi: Closed. open new issue? thank you.
Patrick St-Louis: For mines act permit, old process of vc redistribute. Prob is original docu is long pdf form, so version one of cred is more proof of posession than the permit itself.
Mahmoud Alkhraishi: 931, Rename usda sc-6 form, russell
<tallted_//_ted_thibodeau_(he/him)_(> Guessing we're working from this sort?
Nis Jespersen : Straightforward name change, sc-6 was minimal, now in line
Nis Jespersen : 932, Merging two duplicate credentials
Mahmoud Alkhraishi: Conflict, merge out of band
Nis Jespersen : 933, Two fixes for ag
Nis Jespersen : 944, Renaming organic certification to certificate
Nis Jespersen : Merging quant and measuredval
Ted Thibodeau: Yes but measured indicates it was actual instrument measuring
Mahmoud Alkhraishi: Yeah, different purposes
Nis Jespersen : Addnl prop saying "source", could be "instrument" or "human reading"
Mahmoud Alkhraishi: Sure
Ted Thibodeau: Modify descriptions rather than removing either
Mahmoud Alkhraishi: Term "sensor" / "human user" might be useful, but don't want to go through and make changes
Ted Thibodeau: Close and make an issue. if likely discussion, issue before PR.
Nis Jespersen : Sure, closing
Mahmoud Alkhraishi: [In response to Patrick q] both use spec, how they use not relevant, but
Patrick St-Louis: Workflow instance id belongs to workflow def?
Mahmoud Alkhraishi: Workflow may have many definitions, could say "I'm importing strawberries", 2 forms, avocadoes 3 forms.
Patrick St-Louis: Yes. How about if cases where are diff pipelines, canada, US. Is it expected that they'll also need to become in this trace spec? or ok to have players not in spec?
Mahmoud Alkhraishi: Say gov agency recieving these things. Could say only accept data in if trace. Or could say "I accept trace, but also accept paper". Entirely up to receiving party what they accept.
Patrick St-Louis: Addnl q, entity wants to issue title to title holder, issue it and send presentation to title holder?
Mahmoud Alkhraishi: Yes. How you exchange creds up to you, trace outlines use of presentations. Could do did could do email, but trace creds says "please do this way if you do this way ppl will understand you.
Mahmoud Alkhraishi: 1-1 After call Patrick? Great
Patrick St-Louis: So we do want to use datetime?
Mahmoud Alkhraishi: Ted talked last time abt which interoperable. Can go date to datetime, but can't go datetime to date.
Ted Thibodeau: Which one will hold the value that matters. If need time, need datetime.
Mahmoud Alkhraishi: Ik a lot of these datetime is relevant. Worse case scenario they give datetime of midnight, if don't have more granular.
Patrick St-Louis: I can have a look, don't know if will have time to finish it
Mahmoud Alkhraishi: Did say 1.0, Patrick thanks
Ted Thibodeau: Orie said can close when datetimes addressed, but issue broader. Need to be clear on xsd or schema, but that's just an example.
Mahmoud Alkhraishi: Going through link, trying to find what he's advocating for
Ted Thibodeau: Use datatypes only when that's the only option
Mahmoud Alkhraishi: And a lot of them we could use xsd datetime
Ted Thibodeau: Finding all datatypes not easy
Mahmoud Alkhraishi: What's harm in "not indexed by semantic repositories", which Vlad cites as a reason we should avoid schema.orog?
Ted Thibodeau: If on web, want to be indexed by spiders. Vlad says don't index schema text, do index xsd string.
Mahmoud Alkhraishi: To me q of whether we use datetime v date is 1.0. Question of getting off more generally is not 1.0 I think
Ted Thibodeau: Right call I think, but be clear abt what's happening on this issue.
Mahmoud Alkhraishi: Commenting, we should address date v datetime discrepancies, not looking into removing all for 1.0
Patrick St-Louis: So I should look at vocab, change all date to datetime?
Mahmoud Alkhraishi: If it makes sense. Some places ppl providing date, can't provide datetime.
Patrick St-Louis: So can start by making list, can discuss which are more suitable.
Nis Jespersen : Just use datetime? They can use midnight?
Mahmoud Alkhraishi: But how distinguish bn real midnight
Mahmoud Alkhraishi: But v1 for this PR, switch it all to datetimes, then PR review see if any of these shouldn't be
Mahmoud Alkhraishi: 700, Orie, add section on gs1. Commenting, are you interested, will need it soon for v1. Assigning to Paul, if they're not interested they won't have one
Mahmoud Alkhraishi: 702 Nis our two specs have drifted (interop/vocab). so we need to update cred examples to only use UUIDs.
Patrick St-Louis: Doesn't spec specify needs to be URI?
Mahmoud Alkhraishi: Data model spec says URN, interop says UUID. So are we ok with having any valid URN, or always UUID?
Nis Jespersen : Yeah can prioritize
Mahmoud Alkhraishi: Wait, next issue is your issue on this, with PR that addresses it. Closing
Mahmoud Alkhraishi: Use slash not hash. I feel it's nice to have and huge change, shouldn't block 1.0, if we implement this
Ted Thibodeau: Wait 575 conflation of value, we don't have an issue for steel
Mahmoud Alkhraishi: Reopening, Russell handle steel?
Nis Jespersen : Sure
Mahmoud Alkhraishi: 826, Slash not hash, nice to have, huge, shouldn't block 1.0.
Patrick St-Louis: I know cred id was recent update on [?]. When you assign a cred, should id be provided in request? Or would system generate it and include it in return? We discussed, we said system should not generate its provided in client?
Mahmoud Alkhraishi: In trace side we requested it be provided from client, unless I'm forgetting
Patrick St-Louis: So if id is req, we shouldn't issue cred if not a property?
Mahmoud Alkhraishi: We landed on, credId req in issued cred, not a req on json being sent to be issued
Patrick St-Louis: So do we make it so system can add id if not provided by client?
Mahmoud Alkhraishi: Originally said client should always provide, then ran into issues said as long as issued cred has id should be ok, we shouldn't return 400 if no id.
Patrick St-Louis: So last week there was PR recommending against this, system should *not* gen id. If system is slow doesn't generate response, lost cred in system
Mahmoud Alkhraishi: Valid, but cred IDs relevant, most use cases creds revocable, unusable to have a revocable without id. but bc vast majority are intended to be revocable, we need id.
Patrick St-Louis: They define option for issuance, could have issued cred without id but still refer to it by internal id
Mahmoud Alkhraishi: In interop, says VCs in presentation MUST include ID prop
Patrick St-Louis: Must it also contain status?
<chris_abernethy> apologies folks, I need to drop
Mahmoud Alkhraishi: No, just ID.