[MINUTES] W3C CCG Traceability Call - 2023-04-11

Thanks to Our Robot Overlords and Benjamin Collins for scribing this week!

The transcript for the call is now available here:


Full text of the discussion follows for W3C archival purposes.
Audio of the meeting is available at the following location:


Verifiable Traceability Task Force Transcript for 2023-04-11

  Orie Steele, Mike Prorock, Mahmoud Alkhraishi
  Our Robot Overlords and Benjamin Collins
  Nis Jespersen , TallTed // Ted Thibodeau (he/him) 
  (OpenLinkSw.com), Benjamin Collins, Orie Steele, Russell 
  Hofvendahl, Ted Thibodeau

Our Robot Overlords are scribing.
TallTed_//_Ted_Thibodeau_(he/him)_(OpenLinkSw.com): Recording is 
Benjamin Collins is scribing.
Benjamin Collins:  Okay I'll go and Scribe.
Nis Jespersen : 
Nis Jespersen :  This is a PR for the shift to the non-drafgt 
  version of the UN/CEFACT vocabulary
Nis Jespersen :  This Pull request ended up taking a significant 
  amount of manual work to make sure that all of the terms resolved
Nis Jespersen :  We have two approvals, I'll go ahead and merge
Nis Jespersen : 
Nis Jespersen :  This was the only PR for trace-vocab this week
Nis Jespersen : 
Nis Jespersen : 
Nis Jespersen :  This is on agriculture, we pinged Russel, about 
  half a year ago.
<transcriber> Russell_Hofvendahl_(mesur.io): Yeah this looks 
  doable I guess.
<transcriber> Russell_Hofvendahl_(mesur.io): Allocate some time 
  to catching up in a backlog of old traceability tasks but this 
  continues to look doable.
Russel: This looks doable, I will need to prioritize the backlog 
  to get time for this.
<transcriber> Russell_Hofvendahl_(mesur.io): Yeah that's good.
Benjamin Collins:  What's up next we turn on subtitles.
Nis Jespersen : 
Nis Jespersen :  This is from Mahmoud who is not on today. We 
  want to define what we mean by version 1. How can we progress 
Nis Jespersen : 
Benjamin Collins:  I think we want to make concrete proposals. 
  One would be to have criteria for credentials, freeze the 
  version, and then any changes after that would need to go into 
  version 2.
Orie Steele:  Currently we do build a default vocab-expanded iri 
  for any term not in the v1 context. The format is not very 
  readable. We want to the IRI to be build for undefined terms to 
  looker nice-r.
Orie Steele:  It would be also nice that if you clicked a link 
  that it should provide information that, "this is an undefined 
  term, the issuer should use a better term".
Orie Steele:  The coding side of this is easy once we agree to 
  the data-model and i will link to the vc-data-model stuff that 
  has happend since then
Nis Jespersen : 
Nis Jespersen :  We have three thumbs up on your suggestion. Can 
  we formalize that to say what you said in this comment?
Nis Jespersen : 
Nis Jespersen :  Oh wait, I see it got updated. Any objections to 
  the new recomendation?
Orie Steele:  I didn't intend that url as a recomendation, but an 
  example. We have two options, we can copy off what they're doing 
  in VC v2, or we can give people the ability to distinguish.
Nis Jespersen :  If we go with our own, would you still call it 
  issuer-dependent, or undefined term?
Orie Steele:  I would do "issuer-dependent". Let's replace the 
  current vocab that we're expanding with to the one we know will 
  be in the VC v2 would be the least contentious.
Nis Jespersen :  Would you mind adding that to the issue?
Nis Jespersen :  Any objections to using the VC working group 
Nis Jespersen :  Okay, marking as ready for PR
Nis Jespersen : 
Nis Jespersen :  This is from Vlad and assigned to Mahmoud. This 
  is on EPCIS, Let's ping Mahmoud.
Nis Jespersen : 
Nis Jespersen :  The problem is making it working on the 
  linked-data side. I suggest we close this.
Benjamin Collins:  I don't think we have a concrete proposal on 
  this one.
Orie Steele:  In general we don't generate any JSON-LD for 
  domain, range and inheritance. And we lack the experience to 
  build with these pieces. I think Ted understands domain and 
  range, but I don't see how we're going to make progress with the 
  kind of comments we're seeing on it.
Nis Jespersen :  Let's put a comment to that effect and mark is 
  as pending close.
Nis Jespersen : 
Nis Jespersen :  For this the EPCIS I've put that aside because 
  GS1 came out with their data model, which is more relevant to me. 
  Let me ping myself on this.
Nis Jespersen : 
Nis Jespersen :  This is from me, we're calling ourselves 
  trace-vocab, but we're trying not to define our own vocab. It 
  seems like a misleading name for our project. It causes friction 
  and resistant against trace-vocab like on the UN side. Trying to 
  explain that we don't have not-shipped-here syndrome.
Nis Jespersen :  Does the problem make sense?
Ted Thibodeau:  We're not taking the schemas whole-sale so 
  effectively we're making an ad-hoc vocab, saying these are the 
  vocabs that make sense and using the terms as is.
Orie Steele: +1 Ted
<orie> maybe some more introduction text is necessary
<orie> @nis consider assigning yourself :)
Ted Thibodeau:  We're saying, 'these are the ones to use'. So i 
  think that trace-vocab is the way to go. We can describe our 
  approach to make it more clear.
<orie> again, +1 Ted :)
Nis Jespersen :  I'm happy to close this, but I wanted to bring 
  it up.
Nis Jespersen : 
Benjamin Collins:  I can quickly summarize this. Our schemas are 
  a lot larger than the examples that are being defined. So 
  including Organization makes a schema massive, even though the 
  example is only 50 lines of JSON. So the schema isn't 
  representative of the object that it's trying to define.
Nis Jespersen :  I understand, there's some cases where you can't 
  do that.
Benjamin Collins:  I think that we could have some threshold like 
  90& coverage and then pull that into our CI tools.
Nis Jespersen :  I think we should be getting the requirements 
  down before jumping into tooling.
Benjamin Collins:  I'd be jumping into tooling to break it and 
  see where and how it breaks before following up with the nuance.
Nis Jespersen :  Sounds good marking as read yfor PR and 
  assigning Ben.
Nis Jespersen : 
Benjamin Collins:  It looks like this has a Pull Request that's 
  been merged. Can we mark it as done?
Nis Jespersen :  Looking through the PR, it looks like there 
  aren't any FIBO vocab, which is what I think it is.
Benjamin Collins:  We can try to declare victory and see if 
  anyone takes it away from us.
Nis Jespersen :  Oh, I see it's a schema, but it's redefining all 
  terms. I bet it would be easy to update those terms to point at 
  something real.
Benjamin Collins:  It might be something for the roadmap where we 
  declare a requirement to use certain terms, which would cause us 
  to come back to this.
Nis Jespersen :  I think that might be going too far. There are 
  terms that are not defined. I guess we can close this.
Nis Jespersen :  Never mind, let's mark it as pending close.
Nis Jespersen : 
Nis Jespersen :  Finally something we can close. We just merged 
  my PR for this earlier.
Nis Jespersen : 
Nis Jespersen : 
Benjamin Collins:  This looks like a placeholder to comeback and 
  restore previous changes.
Nis Jespersen :  Let's mark it ready for pr and assign ourselves.
Nis Jespersen :  I found this, we should not have invalid 
  geo-coordinates. We're not using geo-coordinates as often we we 
Nis Jespersen : 
Nis Jespersen :  This one is from Vlad.
Benjamin Collins:  There are some specific suggestions here. 
Benjamin Collins:  I think it might be expensive to map out all 
  of the classes and then flatten the places where it loops back on 
Nis Jespersen :  I'll assign myself and then confirm with Vlad 
  here to see if what i'm suggesting here makes sense.
Nis Jespersen : 
Nis Jespersen :  GS1 uses british english and we want to use US 
  english. It would be nice to have GS1 take a look at this.
Nis Jespersen :  Actually I think we might be able to close this 
  ticket because they took this action else-where.
Nis Jespersen : 
Nis Jespersen :  Never mind, let's close this if we find a 
Benjamin Collins:  It looks like the last action was to add a 
  pending close tag.
Nis Jespersen :  That was half a year ago, let's close it.
Nis Jespersen : 
Benjamin Collins:  I think we can draft out the PR on the issue 
  and if that's okay.
Benjamin Collins:  My thinking is that we would have the JSON 
  field values be correct to a specific region and then have the 
  terms be the generic international version suggested.

Received on Friday, 14 April 2023 03:23:59 UTC