- From: CCG Minutes Bot <minutes@w3c-ccg.org>
- Date: Thu, 06 Jun 2024 19:44:12 +0000
- To: public-credentials@w3.org
Thanks to Our Robot Overlords and Our Robot Overlords and Our Robot Overlords and Our Robot Overlords for scribing this week! The transcript for the call is now available here: https://w3c-ccg.github.io/meetings/2024-06-03-vc-education/ Full text of the discussion follows for W3C archival purposes. Audio of the meeting is available at the following location: https://w3c-ccg.github.io/meetings/2024-06-03-vc-education/audio.ogg A video recording is also available at: https://meet.w3c-ccg.org/archives/w3c-ccg-weekly-2024-06-03.mp4 ---------------------------------------------------------------- VC for Education Task Force Transcript for 2024-06-03 Agenda: https://lists.w3.org/Archives/Public/public-vc-edu/2024Jun/0000.html Organizer: Kerri Lemoie, Simone Ravaioli, Dmitri Zagidulin Scribe: Our Robot Overlords and Our Robot Overlords and Our Robot Overlords and Our Robot Overlords Present: Simone Ravaoli, Ildiko Mazar, Nis Jespersen , PL/T3, Dmitri Zagidulin, Karen O'Donoghue, Hiroyuki Sano, Japan, Stuart Freeman, Andy Griebel, Laura Paglione, Susan Stroud, Eric Shepherd, Deb Everhart, Sharon Leu, James Chartrand, Kayode Ezike, Danny Done, Phil Barker, Jeff O / HumanOS, Alex H, Timg, Naomi, Nate Otto, TallTed // Ted Thibodeau (he/him) (OpenLinkSw.com), David Ward, TimG, Chandi Cumaranatunge <simone_ravaioli> Good morning, good afternoon everyone. This is Simone speaking, today chairing our Monday call with the W3C Verifiable Credentials for Education task force. Today is June the 3rd. Let's just get started with the housekeeping items. IP Note: Anyone can participate in these calls. However, all substantive contributions to any Credentials Community Group Work Items must originate from members of the CCG with full IPR agreements signed. You can find the link to this in the meeting invites. You'll need to have a W3C account, that is also something easy to do online. Moving on to call notes. Please note that these meetings are recorded. We have a robot transcriber and we also make audio and video recordings, that are archived and available online. If you want to speak, please raise your hand, or add yourself to the queue by typing “q+” in the chat. And please be brief if you have a question to ask, we want to maximise the time for conversations. Thank you. Now to Introductions & Reintroductions. If we have any new people in the call, or somebody would like to reintroduce themselves, please feel free to take the floor. If nobody wants to take the floor, we can move on to announcements and reminders. Anyone has events or things that happened that people should be aware of, again, please queue yourself up and share the news with us. Now we can move on to the main agenda topic, that is Learning and Employment Record Resume Standard (LER-RS). Our Robot Overlords are scribing. <ildiko_mazar> FYI, the minutes and an audio recording of everything said on this call are archived at https://w3c-ccg.github.io/meetings/ https://www.thebadgesummit.com/ <pl/t3> Thanks for the link Deb <dmitri_zagidulin> we can hear you, but it sounds like you can't hear us <ildiko_mazar> We can hear you <ildiko_mazar> Yes, we can hear you. <dmitri_zagidulin> we can still hear you Our Robot Overlords are scribing. PL/T3: 1 of which uh some of which are listed here it's tamper evidence it's um certainty of the whole wallet holder's identity I put a big star around that and we'll come come back to that clear descriptions of the claims skills. PL/T3: Structured data link data references to standards and Frameworks uh potentially 4 things like competencies and skills. PL/T3: Of asking for what sought in what scent meaning there's a mechanism by which a relying party can make a request for credentials and the the holder of the credentials can make a decision about whether to share them or not and that's sort of the core set of of attributes. PL/T3: The current state of these vrs is really uh sort of twofold 1 is the ability to send a group or stack of credentials 1 or more in a verifiable presentation to an endpoint such as a credential wallet or some other endpoint. PL/T3: On arrival they are an independent unordered set of in of credentials that is a verifiable credentials the relying party can ha does have agency to choose what gets put into the package that is sent. PL/T3: That's their their agency that they have the actual initiation of the request comes from the relying party themselves. PL/T3: The second alternative a poet for sending credentials um is that 1 can send a compound credential where that is credentials embedded within a credential and that's the model of the CLR V2. PL/T3: The outer credential envelope. PL/T3: Um is a signed document a signed credential which uh has metadata. PL/T3: Then there are embedded credential objects in this case with clrv 2 obv 3s open badge version 3s single assertion credentials. PL/T3: Um they are used in the case of the crv20001. PL/T3: the package. PL/T3: A hole can be signed and the obvious can be left unsigned if the institution wishes that's issuing. PL/T3: Or they can both be signed. PL/T3: Um as as in doing so allowing the individual obvious to be separately verified if they are uh if the individual who was a recipient of them wishes to do so. PL/T3: The composed credentials neither of these approaches. PL/T3: The holder or. PL/T3: Or the school at the job the holder. PL/T3: For either a job or a school application the ability to orchestrate the sequence of their credentials to the relying party highlighting what the holder may want to emphasize and I mentioned that simply because. PL/T3: That's the current state when you get a stack of obvious threes in a VP. PL/T3: Then those credentials are there for the individual relying party to look at as they wish or not in any order they wish or not and that's the same as uh a true of the uh of the compost conventional as well. PL/T3: Why does this matter now. PL/T3: Well because we're increasingly needing to use the capabilities of linked claim of verifiable credentials and Link claims is a particular framework for doing so for a particular population which I'll get to so there's the obvious concern today with respect to the of validity of credentials in terms of their are they actually from somebody that's real um to what extent are these credentials uh fabricated and not associated with the uh an issue or that uh that you can identify in some fashion. PL/T3: Um and most importantly from our perspective. PL/T3: There is an entire segment of the population that is likely not to have issued credentials from an institution or training body. PL/T3: Um or from an employer who has the ability. PL/T3: The present time none are exercising at this point to credential their own workers. PL/T3: So the self-employed the non-traditional workers those who are in small businesses that can't afford the infrastructure necessary to issue and or process verify the credentials um constitute a population for whom our current architecture and design essentially overlooks. PL/T3: So with the interest in skills-based hiring and advancement. PL/T3: We are interested in putting a surgeons of credentials. PL/T3: Endorsements in the same secure verifiable context as assertions of achievements licenses and other uh verifiable credentials like degrees. PL/T3: We want to mirror the pattern of social recommendations currently used for job applications and resumes. PL/T3: Where typically you send in a CV or a or a resume and at the bottom you list 4 or 5. PL/T3: Whom you suggest as recommendations to support the contents of your application and the validity of your of your resume or CV. PL/T3: But we'd like to do that with the same architectural framework that VCS and general possess. PL/T3: And more importantly we'd like to make sure that they're recommending credential is designed specifically for that purpose that is it contains bonafides so that the relying party can judge the recommenders knowledge and the value to uh that the recommender is suggesting in terms of the prediction of the ability of that particular applicant to do the work the job is that the job requires. PL/T3: So to do that we need recommendations that allow you to bind together 2 different verifiable credentials. PL/T3: in a. PL/T3: Way that is. PL/T3: 1 credential to the other that that link between the 2 is tamper evident. PL/T3: We also want to be able to bind in a tamper evident fashion links to digital objects that are somewhere out on the internet for example evidence that you might have and you've placed it in say a Google Drive you'd like a link from your credential to that evidence in that Google Drive that is temporary evident in the sense that since that love that evidence was placed in that Google drive it has not been changed. PL/T3: And we'd like a method to provide confidence in the credential subject that that credential subject particularly the self issuer is who they say they are there's an asterisk by that because that is part of the future work which I'll get to at the very end. PL/T3: The general framework um I'll pause for a second so I can't see speaker notes if there are any questions at this stage. PL/T3: Um or common metric or Karen as to my intro. PL/T3: And I'm hearing nothing. <dmitri_zagidulin> no questions so far Karen O'Donoghue: Nothing for me I thought that was a really good intro thank you. PL/T3: So maybe I should stop sharing. <dmitri_zagidulin> we can hear you though PL/T3: Stop sharing for a second. PL/T3: Can you all was were you able to hear me through that. Dmitri Zagidulin: We were able to hear you. <simone_ravaioli> Yes we heard you. <kayode_ezike> Yes, we can hear you Phil <ildiko_mazar> Can you please explain Phil what the significance is of ordering/sequencing the credentials and what makes it not possible? PL/T3: Okay were there any questions because I'm not hearing audio from anybody at this point. <simone_ravaioli> No questions <karen_passmore> Nothing from me at this time PL/T3: Okay I see a question that eidos mentioned can you please explain what the significance is of ordering and sequencing the credentials and what what uh makes it not possible. PL/T3: Issue is not that they can't be. PL/T3: That an individual can't choose to click through them. PL/T3: Uh as they wish the issue is that the the um. PL/T3: The person that has sent the credentials that is to say the person who is the credential subject the person whom the credentials are about does not have a way for indicating to the relying party the order and the focus of the credentials that they're receiving. PL/T3: That is I can't. PL/T3: The relying party I'd like you to go in this order with these credentials. <ildiko_mazar> Is that a credential FORMAT limitation or a DESIGN feature? PL/T3: Um I can't say that within a given credential I would like to highlight these particular elements potentially relevant in my view to the application that I am applying for or the position I'm applying for ETC none of that is possible I can simply present the credentials in the context of the combined credential uh in the CLR V2 fashion there may be the opportunity for these things to be grouped in a standard way based on some template that represents what we think of as a resume or a CV but I'm not in any way able to point. PL/T3: Sequence of going through those uh to the relying party or say anything about uh the evidence or other elements in those credentials I would like to highlight. PL/T3: I hope that addresses the question. PL/T3: What else do we have here uh. PL/T3: And these questions are coming in faster than I can respond to them. PL/T3: Demitri do you want to take some of these uh because I've got to get my audio fixed because I am hearing nothing. <dmitri_zagidulin> sure Dmitri Zagidulin: Sure yeah so uh the 1 Thing. Dmitri Zagidulin: That I would mention. Dmitri Zagidulin: Uh to iOS follow-up question of is this a credential format limitation or a design feature uh it's basically a a question of whether the issuer arranges it. Dmitri Zagidulin: Or can the receiver arrange it right that's that's the main design question its. PL/T3: Because I question. Dmitri Zagidulin: This is the. Dmitri Zagidulin: This is the general Choice design choice between. Dmitri Zagidulin: Kitchen sink credentials like your full transcript versus. Dmitri Zagidulin: Pending individual credentials of your courses and allowing you to. Dmitri Zagidulin: Um compose them into some of them into resumes uh into other credentials right so it's just. Dmitri Zagidulin: Uh uh single picture versus uh puzzle pieces. PL/T3: And I've got your audio now so I'm you can hear me. <ildiko_mazar> :+1: Dmitri Zagidulin: Oh good good yeah we can hear you welcome back. PL/T3: Okay weird um okay so I will go back is there anything else in the in the question list that that um hasn't been addressed that we need to respond to the deco. PL/T3: Okay so I'll go back to uh presenting. PL/T3: And go there. PL/T3: that's where I. Our Robot Overlords are scribing. <sharon_leu> Can you clarify - in the model with the CLRv2, where it's mixed VC/Non-VC, are links only the VCs or can you link a claim to a non-VC, too? PL/T3: And I would like to. PL/T3: Go off of sharing now and turn it over to Karen to go walk through briefly uh a view of what this might look like. PL/T3: Sure go ahead. Dmitri Zagidulin: I I can I can take that uh and yeah go ahead Karen um if you could share your screen and say a few words of introduction But to answer Sharon uh you can link. Dmitri Zagidulin: Uh the the general notion of hash links which is what which is a link plus a cryptographic hash of the destination can be applied to anything to credentials to PDFs to non VC documents and so on. Karen O'Donoghue: Okay can you all see my screen. Dmitri Zagidulin: Uh not yet. Karen O'Donoghue: Not yet okay great let me try. Dmitri Zagidulin: We can hear you clear though. Karen O'Donoghue: Thank you uh a lot of share screen. Karen O'Donoghue: Are you able to see my screen now. <sharon_leu> @Dmitri, Is that useful if it's not verifiable though? In the context of this group? Dmitri Zagidulin: Uh not yet I think you have to select uh the window or the table that you're sharing. Karen O'Donoghue: Let me see window share uh. Dmitri Zagidulin: And now we can hear you now we can see your screen. Karen O'Donoghue: All right sorry I'm not familiar 1 second go ahead. Dmitri Zagidulin: Okay oh no problem no problem yeah so Sharon uh you follow-up question uh is that useful when it's not verifiable though so the. Our Robot Overlords are scribing. Dmitri Zagidulin: To create a very simple very dumb credential that I read this PDF right that's the that's the self assured we see that I'm that I'm creating. Dmitri Zagidulin: And if. Dmitri Zagidulin: I if I linked to that PDF without uh cryptographic component if I just linked to it all I'm all I'm putting in there is the URL somebody who's hosting that PDF could change it to a completely different contents uh you know after I issue this credential that that I read this thing. Dmitri Zagidulin: Add any sort of cryptographic hash to your link. Dmitri Zagidulin: You can say I read this specific version and if it gets changed afterwards then the signature the uh then the hash won't verify. Dmitri Zagidulin: So that's that's all the assurance that it gives it doesn't hash links don't give Assurance of authorship. Dmitri Zagidulin: Uh you need additional verified credential mechanisms for that so all they do is just lock down the version of the object that you're referring on referring to. Dmitri Zagidulin: Uh over to you Karen. Karen O'Donoghue: Thank you thanks Dmitri so I'm going to show the experience of a user who is interested in creating a self-claimed credential or for a specific skill that they have and I'm going to be showing the mobile version of this experience although these are responsive designs that we've created so they would work across any any type of device and. Dmitri Zagidulin: Karen do you wanna do you want to say a few words of introduction for yourself. Karen O'Donoghue: Oh sure yep thanks okay so my name pardon me right when you asked that of course my throat's all clogged. Karen O'Donoghue: Um my name is Karen Passmore I'm the CEO of predictive ux and I have a background in excuse me designing Enterprise applications and SAS applications and I was working most recently with Arizona State University on their digital wallet for um verifiable credentials which is where I met Demetri and then subsequently got introduced to Phil. Karen O'Donoghue: Pardon my goodness. Karen O'Donoghue: And we've been working together on the user experience of the linked claims app and our approach to this project has been to look at it from an inclusive design perspective to ensure that it's usable by very different types of people across various intersectionalities uh and paying close attention to things like internet connectivity so how can we create an experience where someone might be working on uh either their local machine with intermittent internet connectivity or perhaps they're working in a public environment and they can't save data to their local device so we've taken a lot of that into consideration as we've created these designs and we've also looked outward to research other people have done in this space to learn what we can about people who. Karen O'Donoghue: Might be creating some of these uh claims and our goal is to First create the experience for the user who would claim a skill and then we're going to move on to the user experience for getting a recommendation for a specific skill um but today uh we just have kind of our first iteration or pass at these Concepts and ideas and so as I show this um bear in mind that you know if you want to give some feedback on anything you see here today it would not hurt my feelings at all and whether it's positive or negative uh so we welcome your insights it all just we want to make the best app possible for uh everyone who might want to use it to include you know folks that are going to take this kind of Open Source platform and extended into an application that they can Envision for themselves. Karen O'Donoghue: Uh so to get us started uh this is the kind of landing page that a user would see when they come here to create a self asserted skill credential. Karen O'Donoghue: I've put some kind of marketing material content on here but of course anybody could put anything here it was just to help tell the story of why you might want to create a self asserted skill credential and the idea here is that this user can come here sign up in seconds to capture their skills and this would be a way for them to stand out from other people when they can say I have a verified skill and I potentially have these recommendations behind that skill to help validate it further and the pitch here is that they can now tell a verifiable story and we're suggesting they can capture volunteer work document their on to job training or claim the number of years of experience they have in their field. Karen O'Donoghue: They can also add their proof points request recommendations. Karen O'Donoghue: And then. Karen O'Donoghue: Nice thing here is that the user keeps total control of their data which I think Demetri and both Phil were alluding to in their presentations and you can share only the information you want and you can revoke access at any time. Karen O'Donoghue: So from here if someone wants to get started they can click get started for free. <dmitri_zagidulin> (this will be a free open-source web app, btw -- shown here in mobile browser view) Karen O'Donoghue: And we welcome them and you know we have this sort of friendly user interface like this person Tessa is going to walk them through the experience although we're not using any AI at this stage and we just asked them where they want to save their linked claims and the reason is that we wanted to make sure that people could get started and if they lose internet access their data could be saved somewhere that they choose or they could choose to save their data manually and then come back later because they may not be able to finish this entire thing end to end so for our demo today we're going to allow people to choose Google Drive. Karen O'Donoghue: And then. Karen O'Donoghue: They want to connect to and then from here once they're validated with that account they can continue to. Karen O'Donoghue: Their linked claim. Karen O'Donoghue: So the. Karen O'Donoghue: First thing we want to ask them is we can we're going to be able to get some information from them if they select something like Google Drive so we want to verify that we have the right name displayed here for how they want to present their uh linked claim and they could change this name if they want to but if everything looks good they can click next and then now we want to ask them about the skill that they want to claim so they can fill in a skill name and then they can describe how they earned this skill and we would provide some prompting text here so I have it already filled in with an example but we're mindful of the fact that some people might not know what to put in to these fields so in the screen before this we actually have prompting text and then on the bottom we have the duration and we chose duration specifically because research has shown that when actual dates are put in here it can create a sort of bias against people uh because it might show gaps in between make. Karen O'Donoghue: Maybe their earnings. Karen O'Donoghue: Their employment records so using duration actually decreases that sense of bias and instead focuses on the level of expertise someone has inquired around a specific skill. Karen O'Donoghue: So from here we choose next and they can describe the skill that they have and we do provide prompting text here in a future State iteration of this design we're going to be adding in the opportunity for people to use AI to generate a description here if they want to or they can simply type in their own description. Karen O'Donoghue: And from here they can then choose to add portfolio items and this is an optional step so if they have videos or photographs or um anything that they want to point to online they can add those items here and this would create URLs within their uh linked claim. Karen O'Donoghue: And they can add as many of these as they want so in this example I'm just adding a video of the perfect shot of espresso and I point to the URL of that video and then if I want to add another I can simply tap this link and then it can add more. Karen O'Donoghue: And from. Karen O'Donoghue: Here we asked. Dmitri Zagidulin: Wait wait I want to I want to pause uh super quick on on the screen here uh so this is this is the the 1 of the things that Phil was emphasizing was talking about when creating selfish credentials right. Dmitri Zagidulin: There's a lot of sort of unknown and a lot of questions with regards to selfish or credentials in general like what sort of value do they bring. Dmitri Zagidulin: And 1 of our 1 of our answers to that is twofold. Dmitri Zagidulin: The selfish credentials are as good as 2 things. Dmitri Zagidulin: This is this is what we're we're doing here. <ildiko_mazar> Would this duration setting then result in a static value, that may be disadvantageous with ongoing practiced skills? Dmitri Zagidulin: Take a video of your skill link to a performance appraisal uh to an essay that you wrote to a GitHub repo so this is this is that screen so this is 1 of the 2 sort of main weight bearing. <pl/t3> As we'll mention later we're also looking at the Competency Explorer which is a query interface to a federated set of skill repositories which one, from LinkedClaims, could issue a query to and pull back the structure skill definition from one of the repos in that federated network. <nate_otto> Are you interested in the ability to link to a skills library so (a) the user doesn't have to supply the skill name and description themselves, and (b) it can be the same skill definition already in use or available for use within job descriptions? Dmitri Zagidulin: Uh features of self uh selfish credentials and then the other 1 of is of course recommendations uh which we'll see later on that's it. Karen O'Donoghue: Thank you that. Karen O'Donoghue: That's a great call out. <dmitri_zagidulin> @Nate - very much so! Karen O'Donoghue: So from here after they added their evidence we asked them if they want to add a main image to their credential so this is really just about you know design and formatting and making it present nicely if they want to add an image here and the idea is they could point to really any image that they have or you know they could even go out to something like unsplash and you know grab a an image from their. <dmitri_zagidulin> @Ildiko - GREAT question (with regards to duration vs ongoing skills!) <pl/t3> @Nate - note my call out to the C3 federated repository project. Karen O'Donoghue: And then if they select next they can see the credential that they've created so we have our main image that has been added here um the person's name has claimed basic Barista training uh how many days it took and then the it has their description and then how they earned the skill and a link to any portfolio items and then beneath that we make sure that they're aware that by selecting this box you understand that you are creating a public link to this credential so they know that this is this is now a publicly available link and they can choose not to do that if they want to and at any point during the creation of this credential they can save and exit or they can you know go back to another step and make some edits if they're not happy with the way that this looks but if they are happy with everything they can select finish and sign and it saves it to their Google Drive and so now they've uh saved their achievement and we're exploring whether or not we're going to promote sharing this on social media um so that's that's a conversation that we're having but if we don't the minimum they would see. Karen O'Donoghue: Some visual represent. Karen O'Donoghue: Within this app of the credential that they've created. Karen O'Donoghue: Um when they select this they can copy the link to that credential. Karen O'Donoghue: And then. Karen O'Donoghue: And then as a next step we want to encourage them to go get a verified reference so get a recommendation today and if they select that button it takes them into a flow where we've just provided them with some copy that they can um actually literally copy we've given them some text along with a link to um send someone to a place where they can write up the recommendation which would take them the. Karen O'Donoghue: Recipient back to the linked claims application to write a recommendation that could get then attached to the original credential so the user can choose to copy this text and then paste it into their uh Native email client and so they would copy this text and then um when they're done we would just say oh you've requested a reference now we're not able to track this at this point um. Karen O'Donoghue: I can let. Karen O'Donoghue: To me to speak to that so that's the entire flow that we have so far. Dmitri Zagidulin: Thank you so much Karen so we had a couple of uh really good questions in the chat so ilao brought up a great point about ongoing durations right so what to do when like it's 1 thing if I took a 5-day intensive course in I don't know espresso making and it's another thing when it's I've been practicing this uh since uh 2023 ongoing up till now right and and it's important so 1 that that's a that's a great design consideration for everyone dealing with these skills uh the main thing is that um main thing to note is that that duration skill is free form text right so I can say 2023 till now right I I can indicate or rather I have as much flexibility to indicate that it is an ongoing skill as I do in like a Word document resume. Dmitri Zagidulin: Uh so that's that thought number 1 uh. Dmitri Zagidulin: And then the second question uh was from Nate Auto on whether we're interested in uh sort of integrating with or linking to uh skills repositories since that's 1 of the major. Dmitri Zagidulin: I I don't want to say unsolved but that's the heavier uiux lift. Dmitri Zagidulin: Of how do we. Dmitri Zagidulin: How do we walk in the fine line between making it user-friendly. Dmitri Zagidulin: Right so like. Dmitri Zagidulin: On on the user-friendly access free form text like. Dmitri Zagidulin: How to pour an espresso right like that's that's 1 definition of a scale and then on the other end of the formal spectrum is uh linking into a um Federal. Dmitri Zagidulin: Skill code repository and and say okay this is skill 1 2 3. Dmitri Zagidulin: As codified by this and this directory right so yeah we're definitely interested in um. Dmitri Zagidulin: Linking to an integrating with any sort of existing skill repositories. Dmitri Zagidulin: Uh yeah. Dmitri Zagidulin: Yeah please please yeah mhm. PL/T3: I can jump in here and say that um we actually have or or ranging a call because as you probably know the um C3 community at T at T3 uh has been working for a while now on a Federated skill uh repository network uh with a qu with a user interface query environment uh system called uh competency Explorer is that if I've got that right and um and the question that we're exploring and just and and going to have an a conversation with the developers of that. <nate_otto> That means referencing skills by URLs (not just names and descriptions that are not as suitable for matching). Skills libraries include https://lightcast.io/open-skills and https://www.wgu.edu/lp/general/wgu/skills-library.html and https://www.openrsd.com/ -- the main benefit being that the same skill definition can be used in multiple places. <ildiko_mazar> :+1: Thanks, @Phil. PL/T3: Can we put in a link or uh uh to that particular Federated search uh system from within this app and or provide a UI that can be translated into a query for that particular Network and that we hope will be something that we will be um pursuing actively because you're absolutely right there is a number of different places where we would like the individual to have an option to put in a structured representation that follows some ontology or taxonomy of of skill descriptions that might also be available to people who are issuing credentials from an institution or what have you um to draw from so we definitely interested in doing that and have a conversation scheduled thanks. PL/T3: Okay um at this point. Dmitri Zagidulin: 1 SEC 1 SEC um. Dmitri Zagidulin: I I just. Dmitri Zagidulin: Want to address uh. Dmitri Zagidulin: Karen's comment about being able to track I hold on I'm just reading uh Nate's. Dmitri Zagidulin: Comment in chat referencing skill by URL. Dmitri Zagidulin: Excellent yeah thank uh thanks Nate we'll we'll definitely take a look at uh. Dmitri Zagidulin: Those links. Dmitri Zagidulin: And as as always we'd love to work with you on uh. Dmitri Zagidulin: On any of. Dmitri Zagidulin: That uh real quick uh Karen's um comment about that we're not uh tracking the whether or not. <deb_everhart_(credential_engine)> that can also be done by linking to individual competencies in CTDL, using the CTID as a unique reference Dmitri Zagidulin: Person received and. Dmitri Zagidulin: Uh responded to the recommendation request so 1 of the 1 of the main design choices that uh we we followed closely here is that this is purely. Dmitri Zagidulin: Compliance side application meaning we don't run a server or a database on the back end so. Dmitri Zagidulin: You don't create an account we don't store the credentials for you we don't track the recommendation requests and the responses right so this is maximally privacy preserving meaning when you hit request a recommendation it opens up it's a male tool link it opens up your email client so that you can email your professor or your your friend or whatever whoever you're asking for the recommendation right and similarly uh since we don't access your email inbox we don't know when they reply etc etc now we have a minimal amount of um. Dmitri Zagidulin: Not tracking of um. Dmitri Zagidulin: Statistics but as much as possible this is uh this is meant to be privacy preserving Karen did you have any uh. Dmitri Zagidulin: Any other comments uh UI. Karen O'Donoghue: Well I think that we're prioritizing as you said the. <pl/t3> @Deb - absolutely, and we're very interested in the repositories that are there like CE's. We're just getting to the point in the design where this on the near horizon ;-) Karen O'Donoghue: Ability for uh users to come here and create something and and own and protect their data as much as possible and I think along the lines you know it's a fine line right about how how far we go with that and then where these are experienced may get uh create might create friction right but we're not trying to solve I think it's fair to say we're not trying to solve every problem right now our goal is to get something out there that people can use to create uh a linked claim and then get their recommendations and as we go through this process I think we'll learn a lot especially as we get you know kind of do an initial release or somebody picks up the code and Builds an application with it that will help inform further iterations of the application. Dmitri Zagidulin: Thanks over to you Phil. PL/T3: Yep okay so um if you can release your screen there so I can jump back in there we go. PL/T3: Now head back to link claims I actually think this is the point where we jump um. PL/T3: Not sick uh into um what the an idea of how the Json LD for doing hash links works and I'll bring this up briefly and let turn it to Demetri to go through it sorry there it is. Dmitri Zagidulin: Oh sure yeah uh so. <deb_everhart_(credential_engine)> @PL let us know whatever we can do to help! Dmitri Zagidulin: Uh yeah this just these couple of slides just illustrate the mechanism of uh what 1 of the. Dmitri Zagidulin: Like right so. Dmitri Zagidulin: It is it is a pairing of a URL and a cryptographic hash as as I mentioned to Sharon earlier to lock down the particular version of an object. Dmitri Zagidulin: Uh to to make it um tamper evident. PL/T3: Okay and this is an example of the same idea of in this case linking out to a random digital object in this case a journal a journal article that might be in the person's uh. PL/T3: Might be in a location that they have access to or their own personal repository of of Journal publications of their of their work so this might be the kind of thing that a recommender would do to give Credence to their ability to judge whatever this particular credential is about in this case it's about being able to to Pilot a UAV um and so they've written a paper on it and so this is a hashling to that paper. PL/T3: So moving right along so what's new here um. PL/T3: This is essentially um. PL/T3: Uh nothing terribly new in 1 cents we're taking advantage of the VC data model and the cryptographic of binding approaches including hash links which have been around for a while but not particularly well used. PL/T3: Um there are they do exist in true age credential they do exist in the true pick c2p implementations um which bind in their particular case and this uh images to the camera uh that uh that captured it um and generate a credential as a consequence. PL/T3: What is new is the introduction of hash links for the purposes of this kind of social recommendation and uh evidence uh pointers. PL/T3: Some in summary then we'll just say that link claims is an on-ramp for incorporating unsigned data into complex proofs claims. PL/T3: Uses this hash link mechanism as a Mech as a format for doing so. PL/T3: Tries to mirror the patterns of social recommendations and the like that uh that exists uh today that we've had for a long time. <deb_everhart_(credential_engine)> for anyone who hasn't seen competencies in CTDL, this is an example of a DOL industry framework competency in CTDL https://credentialfinder.org/competency/ce-6aa26646-a955-4ccd-927d-468ac21f5df8 <deb_everhart_(credential_engine)> similar data available from ESCO PL/T3: The other thing I would say is that um it is specifically oriented towards those organiz those individuals that otherwise wouldn't be getting issued credentials from an educational institution training organization Etc and and and gives them a way to create credentials through an interface that they can then store into a digital wallet into a Google Drive uh etc for them to be able to participate in this ecosystem equally with with others. PL/T3: With that said there's a couple pieces of future work we're very interested in and we're doing some staging in the this link claims project 1 of which um is associated with connecting a mobile wallet to a wallet attached storage so that an individual has uh the opportunity to to manage and control the credentials that they have over a lifetime and not just continually have the only location for their storage being their mobile device. PL/T3: Functionalities that by doing that are enabled um 1 might be envisioned in this context author authorization on authentication to credentials in that storage that is available 24 by 7 um and allows access to specific things uh specific individuals or endpoints to look at and things of that sort. PL/T3: The second uh piece um is associated with another new type of credential which we're calling the narrative credential. PL/T3: and in. PL/T3: Ways this is just um. PL/T3: Represent representing the storytelling part that is missing in resume in RSS and and VCS to you as represented by VCS to date and that is literally to tell a story as you would type a story with links in that particular text to the credentials that back up the claims that your story is making and so the simplest way to think about this is kind of a manifest to a a container where all of the different objects in the container are listed as links that can be followed to see what the credential is for that particular object in the container that's in the sort of generic sense in this case it allows you to tell a story. PL/T3: Think makes the collection of your skills relevant and meaningful to the relying party in terms of consideration for that opportunity. PL/T3: And do so by backing it up with data. PL/T3: And uh and so very excited about the possibility of doing that next and lastly the piece that we have sort of danced around um is the idea that the person writing this self issued credential doesn't have the uh aspect of um a an institution or an organization behind an issued credential to give it some degree of credence and and and um and sense of value from a human perspective. PL/T3: Credential issued by XYZ institution is a differently treated than perhaps a credential issued by Phil Long about myself. PL/T3: So the question is what do we do and how do we make sure that it will very least that the credential that I'm issuing about me is actually from me. PL/T3: And the first step in that is to make a a process that allows you to validate that and do the kyc which issuing institutions do by default when someone applies um to be a student there or in an employment context when someone applies to be a an employee at a censure or some place like that those individuals go through a kyc process and know your customer if you will process to make sure that they are who they say they are. PL/T3: In our case the idea for doing that here is with a notoriousedc. PL/T3: and liter. PL/T3: Leveraging motorization services that exist and which almost every state in the US at least now permits in terms of a remote notary service to allow the individuals identifier in this case there did to be connected to a person and the notary is a new credential subject uh type. PL/T3: that is. PL/T3: Bound 2 with link claims to the individuals credentials that is making their claim. PL/T3: so that. PL/T3: The context of the notary as a future piece of work. PL/T3: Lastly there are lots of other places that this idea can can be applied to uh it can be applied to for example uh keeping track of the papers you should submit if you're an academic or you're writing you're a business person you want to have a repository of the things you published it can be used to do reputation reviews it can be used to as you saw in my reference to the c2p project it can be used to um put a credential around the origin of an image and the bonus the the data around its capture to try to um deal with uh the proliferation of of image manipulation that's out there it can be used to track Disaster Recovery funds to see whether or not they actually got distributed to the individuals that claim uh that they were claimed to be sent uh to to help them and others. PL/T3: So that's linked claims. PL/T3: And I'll stop there and we'll open it to any questions that people might have. <ildiko_mazar> Could you share the slides, please? PL/T3: Have good question and title timely because uh I believe we have in fact been able to secure a small amount of funding for doing some usability research Karen would you give a brief uh summary of what your intentions are here just to do this. Simone Ravaoli: https://docs.google.com/presentation/d/1b4iyZi2b2fZ-m65nJR36MUlha419LI898GsjHHZTfVc/edit?usp=sharing <ildiko_mazar> Thank you, Simone. Karen O'Donoghue: Yes absolutely that I appreciate the question as well it's super important to me to do the not only usability research but also understand the people who would be creating this data and then the consumers of the data for right now what we're going to be focused on is. <pl/t3> @Ildiko - yes -Simone beat me to it! Karen O'Donoghue: Potential users of that application from the perspective of creating a linked claim and we're going to take the application that you just saw today and then approached anyone that we are able to recruit um for the amount of time that we have available we're looking at doing maybe a 6 week usability study have them walk through the application with us uh through some various scenarios and get their feedback on the accessibility usability um intuitiveness of the application so just making sure that we're actually creating something that people can use another part of that that we'd love to embark on is doing some deeper research and understanding when people might want to create a linked claim and then how they want to use it right so we're showing kind of a very limited subset of the possible use cases that fill displayed on 1 of the last slides there and we would love an opportunity to go deeper into that research and also understand um things from different perspectives so we have you know traditional Learners and non-traditional Learners and people who are maybe going to be creating. Karen O'Donoghue: A linked claim who might. Karen O'Donoghue: Currently be employed for. Karen O'Donoghue: To somebody who isn't employed and just coming at it from um understanding those various perspectives would is important to us as well as looking at it from the notary perspective so when somebody is going to provide a recommendation to someone it's important to us to understand you know what they go through this process what kind of blockers might prevent them from making a recommendation because we we know some upfront friction is necessary to create a verified experience here or verified link claim but we don't want to go too far right so then beyond that we're also interested in understanding how companies might consume a linked claim because we understand that you know right now we're at the early stages of Designing this and we've had internal conversations about things like uh Phil mentioned earlier like taxonomies and ontologies and other systems that we could connect to but we also understand that it's important for us to look at the companies that might be consuming length claims and what systems um. Karen O'Donoghue: They're using and how. Karen O'Donoghue: So there's just so much research here that we're excited about doing but we're we're definitely at the early early stages so it's focused on the usability of the current APP you saw today. PL/T3: Last comment on that um we have we didn't describe it here but we do have an interest in being able to present a view of credentials to an individual who does not have a credential wallet that is to selling party that does not have a credential wallet but could be viewing them uh via their browser and verifying them via their browser that does not mean that the credential is sent to them it just means that they are given a link to where a credential is temporarily stored to do that. PL/T3: and then the. <ildiko_mazar> Have you considered a use case for prior learning recognition? PL/T3: Sometimes if they wish to receive the actual data to put into their systems um how would we do that because of the challenges associated with establishing the current mechanisms and protocol exchange for credentials that small businesses are unlikely to be able to support and maintain themselves and we have some ideas about going about that process uh lastly I say the T3 has been generous and the funding is allowing us to put this application up when it's complete for probably at least a 3 to 6 month period for anybody to use and allow us to collect data potentially on its usage and allow others to try it out and find areas that we can improve or that they can improve because it will be open source. Dmitri Zagidulin: Yeah to add a couple uh just a few more words to that uh as as as Phil mentioned uh it's going to be an open source app right it is it is already uh and it'll be that much more useful uh when it's finished so you you'll be able to take a look at it hosts your own. Dmitri Zagidulin: In term will. Dmitri Zagidulin: Will publish the uh credential schemas of of the credentials that were that were creating both the. Dmitri Zagidulin: Uh recommendation ones and the lightweight uh. Dmitri Zagidulin: Skill credential so. Dmitri Zagidulin: And and we were definitely uh there's lots of exciting usability research especially as both Phil and Karen mentioned in the. Dmitri Zagidulin: The receiving relying party end right if all this is successful and I get a self self asserted credential uh with recommendations with evidence all this good stuff. Dmitri Zagidulin: And I sent it to my employer how will the employer consume it how will they display it how will they be able to verify it not having verify your software of their own and how do they uh work it into their uh existing workflow all of those things uh we're looking at so please uh please come talk to us to Convention have we considered a use case for prior learning recognition um so it's a 2 things 1 is I I think other projects are focusing on that a bit more right so the traditional uh learning recognition. Dmitri Zagidulin: Credentials but uh the the other answer is the. Dmitri Zagidulin: Self asserted uh learning and skill credentials that that we're doing in the app right now are flexible enough to recognize those as well uh but we'd love to talk to you more about it. Dmitri Zagidulin: To see what you have in mind. <pl/t3> ACE has a prior recognition for credit initiative. LC's could be adapted for that. Dmitri Zagidulin: Any other question. PL/T3: Over to you Demetri. PL/T3: Oh excuse me over to you Simone. <ildiko_mazar> :+1: Thanks @Dmitri. Dmitri Zagidulin: Yeah uh any anyone else Simone okay yeah. <deb_everhart_(credential_engine)> see ACE credit recommendations in CTDL https://credentialfinder.org/search?searchType=competencyframework <deb_everhart_(credential_engine)> linked data is useful for connecting to other resources <sharon_leu> Great work! <deb_everhart_(credential_engine)> ditto, great work all, thank you! <jeff_o_/_humanos> Great Demo, thx! <ildiko_mazar> Great work, great demo, thank you.
Received on Thursday, 6 June 2024 19:44:20 UTC