Verifiable Claims Working Group

06 Jun 2017


See also: IRC log


Dan_Burnett, Richard_Varn, John_Tibbetts, Liam_Quin, Gregg_Kellogg, Manu_Sporny, Dave_Longley, Chris_Webber, Christopher_Allen, Adam_Migus, Matt_Stone, Colleen, Charles_Engelke, Joe_Andrieu, Matt_Larson, Ted_Thibodeau, Drummond_Reed
RichardV, MattS, DanB


<Colleen> +present

<scribe> scribenick: cwebber2

<burn> scribe: cwebber2

richard: any changes to the agenda before we start?

gkellogg: have a statement before doing the publishing of the use cases, only open issue I know of is the mailing list for the process. We've got the working group mailing list, but we still wanted to do another mailing list?

richard: Okay we'll add that to the agenda, when did you want to add that?

burn: it should go with the discussion of the publication of the document

richard: ok
... any other additions/changes for the agenda?
... anyone new to the group who needs an introduction?

<stonematt> link to agenda for this week: https://lists.w3.org/Archives/Public/public-vc-wg/2017Jun/0003.html

tallted: Hi, I was just nominated from my AC rep, I have some knowledge of VC... I'm from Virtuoso, which is one of the primere rdf projects. we're also doing things with openid webid and etc

<Zakim> manu, you wanted to discuss recruiting...

<ChristopherA> (don't forget re-introductions)

manu: just a real short addition to the agenda, we may want to discuss recruiting more folks into the VCWG. we want a regular heartbeat on getting new folks into the group. usually you see a rush of people in the first 3 months or so as it becomes official

richard: add that to the standing agenda or?

manu: let's do it at the end

<Varn> fq?

<ChristopherA> (we have re-introductions on our standing agenda now at CG)

richard: anything else?

<ChristopherA> Maybe a quick text intro?

stonematt: This is Matt Stone, I'm with Pearson. We're with the ??? division. I'm one of the chairs of the WG, my background goes back 15 years or so as we built a commercial platform built to manage and do data management for professional credentialing, eg in the IT space, nursing, etc. Part of our goal all along was to have professionals share their credentials with employers etc

<stonematt> Pearson VUE

richard: our status of publishing, we basically agreed we were going to publish this as a note
... so aside from the question about the mailing list, anything else that needs to be covered?
... anyone have any questions or comments on this agenda item?

<Varn> ack: liam

liam: it turned out that we were going to publish it as a FPWD to go to rec to get IPR requirements. but turns out since we're publishing it as a note our IPR requirements don't apply. We won't get any additional protection from that. If we get it in today we can publish on thursday or next week this time

<manu> +1 to publishing on Thursday

richard: let's go back to the mailing list question

<Varn> Liam--can you comment on this as to what W3C would prefer

gkellogg: So, traditionally the working groups have had a few different mailing lists. we have a working group and members mailing list, one is publicly visible by members and one is private. There's also traditionally a third comments mailing list. that's typically how it'd be used for specs like this to solicit input, etc. that's so conversation about comments etc don't bleed in to other group activity. so we could change the

status to vc-comments@w3.org or whatever, but it might not make sense to set that up if the ML doesn't exist

liam: process requires a mailing list, I'll get one created to the maling list, so we should say you can publish to public-vc-comments, but it's common to say that issues should be filed on github, but if you can't do that for whatever reason you can use the mailing list
... that way everyone can see there's a summary for all your issues
... the one thing you mustn't end up with is a place where you have issues all over the place

richard: github will be the official place, can be commented elsewhere, but eventually things need to end up on GH

gkellogg: in that regard I think adding a sentence to the status of this document indicating that comments can go either to GH or the mailing list would work. We're limited to what we can do there, so it will have to either go to the beginning, the second paragraph, or to the end there. Presuming that's what the group wants to do I'll also add an issue about being able to file issues on github

richard: ok... anyone have an objection to what gkellogg just laid out?
... looks like no, so gkellogg go ahead with documenting that

Discuss FPWD for Data Model doc--what issues are blocking finalizing the FPWD and address

richard: anything else on agenda item 2? no? ok let's move on to item 3, which is the data model doc
... we looked at this the week before, we agreed this is the main thing we wanted to focus on. There's some that have been out there that have been dated. I think there were like 43, or there was a chunk of them out there, we don't necessarily know what'd keeping them from moving to resolution. so we should find out what to spend time focusing on

<Zakim> manu, you wanted to provide background on this item

manu: so it's true we have around 35 issues right now, but I think there are really only 4 to get vc models to FPWD. Good news is I think we've made solid progress on 3 out of 4 over the last week. before diving into each one of those items, what i think I'm asking for is for the group to sign off on integrating those into the spec so we can tackle those before the call next week. for background I've been acting editor for last week

or so trying to get these issues resolved, has been good discussion between a number of us on these threads.

<stonematt> manu: are you talking about the PRs?

manu: at a high level we're checking if there are anything anyone has to comment on the changes on these specs. I'll go through each one, asking if there are objections before we go forward with merging them into the spec

<stonematt> thanks.

<manu> https://github.com/w3c/vc-data-model/pulls

manu: yes I'm discussing PRs.
... these are the current pull requests active on the document; the fourth I can't really do until the group approves/denies. the fourth requires pulling them into the spec. this will help me make changes to the fourth one
... let's go through one at a time
... easiest first

<manu> Add content to Principle of Minimum Disclosure section. - https://github.com/w3c/vc-data-model/pull/38

<amigus> I just sent you the patch

manu: here's the first one, this one's about adding principle of minimal disclosure section to the spec. this is a holdover from the CG days. joe and adam were looking, joe got his comments in, but adam, I don't know what your comments were
... adam, maybe if you could... ok, adam sent me the patch I'll take a look
... adam / amigus, could you look at the differences in the patch I sent

amigus: okay, yes, so it's really just word shifting, taking what Joe did and shifting it based on my own meeting. you can look if you like my wording better, and if so merge it, but otherwise there are no functional changes

manu: ok I'll look at the diff and try to integrate, is that ok?

amigus: sounds good

manu: any objections to merging this into the document at this point?

richard: not hearing any objections, so pull that into the spec

<manu> https://github.com/w3c/vc-data-model/

manu: next one... just a quick mention, some of the people on the call may be a bit lost at this point, like why wasn't I notified about the PR. if you weren't notified, you're not subscribed. you can go to this link and click "watch" to get notifications. if you aren't watching you won't get notifications.
... ok next up is commenting up webidl section

<manu> Comment out WebIDL section - https://github.com/w3c/vc-data-model/pull/45

manu: so this pull request is about commenting out the webidl section. it doesn't remove it, it just prevents it from being displayed
... one concern is that webidl adds a lot of complexity that we might not want to add at this point in things.... we want to make it clear we're putting out a simple system

<Varn> turns it into a comment so it is not displayed

manu: second is webidl was throwing errors
... third is people may assume this requires browser integration, and we've also noticed in the web payments group that adding webidl means that browsers start adding strict analysis. we're not saying we're not going to do webidl, just commenting it out, do fpwd, and leaving it for later

stonematt: one question for this in relation to fpwd is one thing we said in previous meetings is even if we include content in the document at this point of publishing we would/should include sections indicating we would follow up in areas. is webidl one of these? are we completely silent on it?

manu: we're presently silent on it, but we could add a marker saying we're contemplating adding it and we're seeking input

richard: any preference on which we do

manu: I think it's reasonable to add an issue marker

dlongley: I'm also fine with an issue marker, we could make it more general making sure we have a number of representations

richard: ok proposal is to add issue marker saying we're considering other... what's the wording?

manu: editors will work on the wording don't worry

richard: ok including but not limited to webidl, other things will be looked at... anyone object to this resolution to pull request 45?

<burn> ACTION: editors to add issue marker for new sections that might be added to the spec such as webidl [recorded in http://www.w3.org/2017/06/06-vcwg-minutes.html#action01]

richard: hearing none we'll have that as a resolution with editors adding that as the issue marker

<manu> Added this comment: https://github.com/w3c/vc-data-model/pull/45#issuecomment-306523870

<manu> Refactor some uses of "identity" to "entity" - https://github.com/w3c/vc-data-model/pull/49

<dlongley> Zakim: who's on the phone?

manu: ok so, last one is Harriet's one in this group, but I think Joe and I think... this pull request has to do with refactor moving "identity" to "entity". to be clear it's not "we should never mention identity", it's a terminology issue and how we use it in the spec, about how to align
... joe started looking at commenting on that

<Varn> Harriest, not Harriet's meaning troublesome

manu: so we use that as a way to clean up the glossary. to be clear the glossary includes more terms than we pull in, so the glossary has a number of terms we won't include when we publish, so we don't need to painstakingly pour over those... that said joe and I went back and forth on a number of definitions, I think I pulled in all Joe's suggestions, I think there's been more alignment on what we do and don't mean
... Joe, I'd like to hear what we should do before we pull this in?

joe: so I really like the changes from identity profile -> entity profile, all feels good. two outstanding things maybe we should iterate on more. I'd like to remove identity altogether. there's the term "identity provider" and that's out in the field and I don't mind that so much, but I don't like the definition itself and I'd like to avoid the whole debate over what's identity. might be a way to sidestep that rathole
... I've been involved in this conversation with this group for a year now. I only now realized we discuss subjects, statements, claims, those are different terms than what's in RDF

<ChristopherA> I would like to see that added as an issue.

joe: not sure what the right steps are, but we should look at how we're working with people who would see those

<manu> ChristopherA - just add it as an issue in Github :)

<Zakim> manu, you wanted to discuss removing "identity" altogether... and "identity provider".

joe: so in RDF you can think of it at triples/quads, but a given rdf statement is a triple, subject predicate object, so we have definitions that a claim is a statment about a subject, but for example 7 different triples, so colloquially it's a ???

<burn> I think this discussion is no longer minor and has moved beyond the original PR

manu: joe mentioned two things, removing identity altogether, and the identity-provider term. there are a few examples of identity that would be hard to revove. we could certainly remove from the terminology, but I don't think we pull in identity provider into the...

joe: the terms and glossary has it as an entry

<burn> Not to say that it is unimportant, just that an attempt to get quick approval in this call is not appropriate.

<dlongley> in the statement "Jane likes cats", "Jane" is the RDF subject, "likes" is the RDF predicate, and "cats" is the RDF object

manu: yes but not all of those are pulled into the spec, which achieves partially what you want to achieve. the other thing aboud identity-provider is it's a holdover from the spec of 3 or 4 years ago. then we switched to "credential repository" but I don't think we ever finalized what term that is

<Varn> subject object predicate triplet is standard RDL but we have used them in combination so that we do not use subject in the same way

manu: in any case I don't think it shows up in the spec, it depends on how much of the ecosystem we want to describe
... this will show up in the fourth and final part of the spec
... all that said, more or less what you said is more or less achievable except maybe I heard some pushback on removal of it entirely because may push the wrong message

<dlongley> in our data model, we have an "entity credential" (an RDF subject) that makes a "claim" (RDF predicate) about an entity (RDF object)

<dlongley> the entity is the "Subject" in our claims terminology.

manu: what we may want to do is put an issue marker in the spec that says "yes we know VC can be used for identity, the group is currently set to define it properly so that it's not jabbing other groups in the eye effectively"
... so we may want to put an issue marker in there, to make it clear we're aware and are considering it, leave it at that for FPWD
... on last item we mentioned, I'd like to stay pretty far away from mentioning rdf concepts

<dlongley> +1 RDF mapping is secondary

manu: talking about low-level technical bits of rdf will impact the group negatively in a bad away... I'd rather not mention RDF but talk about it in a bit more high level sense in the data model. that may confuse people who know about RDF but hopefully they know how to map

gkellogg: my concern about subject and statement was to say people may be confused... but these are such broad terms that anyone needs to consider the context in which they're defined
... we do define what a subject is, I think that makes it unambiguous. we don't define object, statement only used in describing a claim, no other discussion of statment, the claim is a statement in our case. if we use terms we define, they should come from our terminology. if we use terms defined elsewhere, we can define it in our terminology and cite the original definition

<manu> +1 to everything gkellogg just said

<Zakim> burn, you wanted to say that we are going beyond original PR and to say that this PR and discussion may not be baked enough yet.

gkellogg: eg we could reference the rdf concepts, which maybe we shouldn't do but we can do a few things to make sure the terms are clear

<dlongley> +1 to gregg as well

<dlongley> +1 to focus on identity => entity conversion

<stonematt> +1 to separate concerns

burn: mainly I wanted to point out that our discussion about subject and etc is going beyond the original PR. I'm uncomfortable going beyond the original scope of the PR, which is about identity and entity. the other topic about subject is important but should show up as a separate issue

<Zakim> manu, you wanted to ask Joe what changes he wants to see before we pull in the PR

manu: just to circle back I need a list of edits I need to make to this PR before we pull it in, so joe what changes do you want to have made before we pull this in

joe: if we're going to keep identity as an entry, let's have another definition... I can propose something... I'll try to get something tonight. It may be too complicated but let's try to evaluate it
... on the topic of the subject-predicate I agree it's not part of this PR and I think maybe it's confusing and that's just part of how we have to do it. that's the only issue I'd say hey maybe we have to work on

manu: there's one other thing about referencing multiple subjects, typically it's about one subject but usually multiple subjects

<TallTed> +1 subject/predicate/object terminology SHOULD be brought into sync, not conflict, with RDF

joe: I think that's a non-issue. I created that issue as in terms of how we define subject
... it's tautological, the claim is not a subject

manu: I don't know if that's true... I think you're right but not for the reasons you outlined

dlongley: I think that's likely a corner case. I think we should put an issue marker in the spec

manu: I think ok, we'll put an issue marker bout not doing SPO, an issue marker about multiple ??? and thn we're done

stonematt: no objections, I pause before I ask the next question

<manu> https://github.com/w3c/vc-data-model/pull/49#issuecomment-306529073

<manu> Make the following changes:

<manu> If we keep identity as an entry, different definition.

<manu> Do not do anything w/ subject/predicate/object.

<manu> Put issue marker about multiple subjects in claims.

richard: any other comments/objections on 49?

stonematt: my question is about other items related to the FPWD as we're closing in on PRs... I notice of the issues you have open, none are tagged with FPWD or labeled with milestone
... just wanted to validate that's correct

<Zakim> manu, you wanted to note fourth and final item to discuss.

manu: they are tagged if you look at the PRs
... oh I see what you're changing

stonematt: of the open issues we have tags for privacy, security, etc
... are any of these worth discussing before we go to FPWD?

burn: I went and marked all of them with the FPWD milestone, I agree that's the better way to do it

<Zakim> manu, you wanted to discuss fourth and final item.

manu: there's only one remaining thing, which is cleanup of the language to make it easy to read, put in diagrams of the issues etc, where we expect things to flow, so that's the fourth and final item to get the terminology down so it's easy for people to read, esp people who are not involved, get a sense of the ecosystem we're designing etc. that will come in a separate PR... probably can't land it till saturday/sunday, which means
... next monday people can review. I think that's the last thing we wanted to do before going to FPWD on the data model spec

stonematt: that gives me an idea of what's next
... so we expect with these PRs that we go through a few rounds before we go to the final...

manu: yes, ??? a few weeks from now, with PR in by next tuesday

<manu> absolutely! :)

<stonematt> +1 ChristopherA

burn: yes all I'm going to say is to remind people that manu is not the only one who can say what they'd like to see in the FPWD. these are just items they brought up. if there's anything others want to see before putting it in the FPWD please file and label it as a FPWD issue

<Zakim> ChristopherA, you wanted to ask the chairs remind the members here of the Credentials Community meeting happening immediately after at 12pm ET 9am PT. (I have to leave)

ChristopherA: I don't want to add to the expanding agenda, but want to remind that the community meeting is right after. I have to go to that meeting now! so that was basically it.

<manu> Agenda and dial in info for CG meeting after this meeting: https://lists.w3.org/Archives/Public/public-credentials/2017Jun/0058.html

richard: well we finished the one we wanted to spend most of the time on, which is good, we don't have time to review the issues for data model document
... any other suggestions?
... for next week's agenda?

stonematt: I like the idea of re-introducing people over the next few weeks

<Drummond> I like that idea - I'm getting re-engaged in VC work

burn: just a reminder, I'm putting the Zakim request. it lists everyone who's present. if you don't see yourself there, please do "present+ <YOUR_NAME_HERE>". please remember to do that before the call ends.
... if you already show up you don't have to

Drummond: I spent the first 30 minutes trying to figure out how to make myself present, and I've done a lot of time on community group type things. the accessibility is near zero

<Varn> IRC is not intuitive

<manu> +1 to everything Drummond just said.

richard: I appreciate that comment, chairs will take this on as a step

Drummond: there's also no information, most groups publish a page of information on how to join a call

burn: the agenda email each week actually says here's how you connect for WebEx, etc

<Drummond> Sorry if I'm venting about how hard it was to join - it may be that I'm in fact not on the right mailing list. My bad.

ChristopherA: the private email you link to should have the URL which is the full url including the password etc. there is a way to do that with WebEx

<burn> Drummond, you are right.

<stonematt> +1 to having a 1-click join from ChristopherA when on the member archive

<burn> As Richard says, some of the indirection is due to W3C policy to avoid abuse

<TallTed> TallTed: there is a double-redirect right now; the password email pointer in the agenda goes to a public message which links to the private message

richard: there's a security issue of not releasing the password, but we have to be careful
... meeting over, thanks everyone

<stonematt> bye all

<burn> scribe: cwebber2

<burn> rrasgent, draft minutes

<burn> Title: Verifiable Claims Working Group

Summary of Action Items

[NEW] ACTION: editors to add issue marker for new sections that might be added to the spec such as webidl [recorded in http://www.w3.org/2017/06/06-vcwg-minutes.html#action01]

Summary of Resolutions

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.152 (CVS log)
$Date: 2017/06/06 16:13:03 $