Manu Sporny: All right welcome everyone to the August 30th 2020 to VC API call. I will put our agenda into chat here and there on the agenda today we will review the doodle poll the new time we also have introduction. ✪
Manu Sporny: Updates in today's basically talking about the VC API so the verifiable credentials API. ✪
Manu Sporny: In the goals design the the status of the suite in the roadmap and then just a general discussion on what people would like to see differently you know happening there and then we might end early or we might try and process some issues so we all take that as as it comes any other updates or changes to the agenda before we get started. ✪
Manu Sporny: Okay if not let's go ahead and jump into it. ✪
Topic: Introductions, Relevant Community Updates
Manu Sporny: Introductions relevant Community updates I think we all know each other here is there anyone that would like to introduce or reintroduce themselves. ✪
Manu Sporny: Okay any Community updates relevant Community updates. ✪
Patrick St-Louis: Yeah I just wanted to mention that tomorrow there's an entire of patent presented by in d.c. oh in the hyper Ledger areas space and as we're going to be hosting a room helping people sort of Test Mobile Wallet application so how to use the area's agent test harness to test Mobile Wallet application and there is a small component about verifiable credential in this so. ✪
Patrick St-Louis: I thought it would be interesting to mention. ✪
Manu Sporny: Yeah interesting do you have a link to that that we could put in the minutes to share with the people. ✪
Patrick St-Louis: I will definitely share link during the meeting I just need to go find it in my emails yeah. ✪
Manu Sporny: Okay great thank you in Patrick would you mind giving us kind of an idea of like what are the goals of this particular interop Athan interop Fest. ✪
Patrick St-Louis: So it's the second interpret on that in this year's hosting the Austin won in the past there's a few goals so at the moment a big conversation and the hyper Ledger in the area space is out of bound connections between agents and there is a move towards a IP version 2 so AI peas are sort of what defines. ✪
Patrick St-Louis: Concepts and Airy space and migrating to this new VIP allows for more features to be present so it's a way for vendors and organization to come together and see if they're updated applications can talk to one another so they will be like an infrastructure in place they will simulate a scenario with laboratory offering test results this is going to be a holder agents. ✪
Patrick St-Louis: Verify agents and it's just basically. ✪
Patrick St-Louis: Showing up and testing specific rfc's to see if they can communicate in between every agents there's for more details I'll give a I can definitely post the links for more details about specifically which rfc's are the main focus of the event but it's mostly for the community to come together and test their limitations. ✪
Manu Sporny: Excellent thank you Patrick and yeah looking forward to that link as well any other relevant Community updates. ✪
Manu Sporny: Okay some of the community updates will cover during the test Suite goals design roadmap B the only thing that may be relevant to some people here working with verifiable credentials let me share my screen in this is a super rough demo playground no Jason they'll be so the Jason. ✪
Manu Sporny: Um has been upgraded to support safe mode in json-ld J so one of the things when you're creating a verifiable credential that most json-ld processors used to do was that they would silently drop information that didn't have a term definition in signed the stuff that that didn't happen and that's been a couple of people over the years and so we've introduced at least in one of the. ✪
Manu Sporny: Mode that will warn you when you're trying to use something that doesn't have a that doesn't have a term defined so if I go to the options here and see the safe mode here it's set to default I forget what the default is but if we set safe mode to True which was what we would expect to be set for most verifiable credential libraries. ✪
<dave_longley> the default is "true" for canonicalization
Manu Sporny: Forget if the activity streams context has a. ✪
<dave_longley> (for safe mode)
Manu Sporny: Hey vocab setting through well we'll see or yeah it does what give me let me use a different example here not schema.org this one if we add an undefined property in here like Foo bar in the json-ld processor detects that it's going to drop a property it will now throw an error so this is very programme. ✪
Manu Sporny: Exposed in this way in the signing libraries but we actually do detect it in throw a warning stating that this is an invalid property in stating that you know it's going to drop the property go ahead Dave. ✪
Dave Longley: Yeah just want to make a quick comment that when you use the canonized tab on the playground the default for safe mode is true and canonized as what ends up getting used in signatures so that with the latest version of the library that fails it fails what does what's called failing closed it defaults to safe mode by true setting that to true so you'd have to turn that off if you wanted to allow unsafe values. ✪
Manu Sporny: So in theory if I do food this it's going to throw an error safe mode validation error. ✪
Dave Longley: The other thing to note here is that we'll also catch relative URLs so if if you're using relative URLs in your data and you don't have any mapping back to the vocab or at base and you have a relative URL that will also catch those errors so one of the most common mistakes people make is they they put at type or type if they have an alias and they put a value in there that doesn't actually map. ✪
Dave Longley: You can see that at that are happening and you know you the same thing would happen if you put a capital f for Foo and no / because it attempts to yeah that when you do that I'm json-ld it attempts to make it a full URL but since safe mode will catch that for you now and throw an error. ✪
Manu Sporny: Well in any case if this doesn't make any sense to you it may not affect you but if it does make sense to you you we are trying to protect developers from accidentally signing data that they they think they're signing when they're really not okay so that's json-ld safe mode will make an announcement on that about that on the mailing list later this week. ✪
Manu Sporny: Okay next up is possible new time for me see API calls question mark we have a dual Pol out because the number of people have said that they've wanted to come to this meeting but it's a bad time so we sent out a pole and we've gotten a lot of responses so far which is great the not sag. ✪
Manu Sporny: UPS in everyone's schedule and it's not clear how many of these people are actually going to be able to show up. ✪
Manu Sporny: We have some good kind of times so I think what we're going to try and do because you know no no time choosing processes is completely Fair we're probably going to optimize for implementers to get as many implementers on this call as possible implementers and potential implementers so they will get first priority and then second priority is potentially implementers and in all other people right so we're trying to just get as many people on these calls has as we can. ✪
Manu Sporny: We do have some times that are looking pretty. ✪
Manu Sporny: It's read 3 to 4 p.m. so an hour before this time maybe in any case go ahead and get your you know thoughts in on their understanding that it's just really hard to find good times for everyone any questions on the new time or any of that. ✪
Manu Sporny: The other heads up I think associated with the time is this month sorry in September it's going to be a heavy conference season for everyone and travel season for everyone w3c TPAC is the second week of September we will probably cancel this call during the second week and then rebooting the web of trust is the last week in September and so we will probably cancel this call. ✪
Manu Sporny: All then as well because a number of us are going to be. ✪
Manu Sporny: Okay anything else on scheduling timing any of that stuff. ✪
Manu Sporny: Oh it's it's well it's Eastern Standard but I think you can set your doodle preferences so it'll put it in your local time time zone. ✪
Manu Sporny: Okay yeah it's it's yeah unfortunately looked it's complicated right yeah it's not it's not supposed to be in any particular time zone people are just supposed to pick it typically what we've noticed over the past 15-20 years is that you either have a lot of asia-pacific folks in a meeting like the internet of things and web of things stuff and then you tend to have it in in daytime. ✪
Manu Sporny: Or you end up having something that's pretty heavily focused Western world's so it's East Coast 10 a.m. 11 a.m. so year up in the u.s. meats and there is very rarely a time that you know well it's just not a time that works for everyone on the in the globe so we'll do our best we I think we optimized to try and make sure that you know matter could participate in these calls in we're going to you know. ✪
Manu Sporny: Continue to see if we can we can do that. ✪
Manu Sporny: The problem okay anything else on the times before we move on. ✪
Topic: VC API Test Suite Goals
Manu Sporny: Okay next up is BC API test Suite goals. ✪
Manu Sporny: I'm going to I'm going to start off please other people join in because this is supposed to be the people's test Suite right our Collective test suite and so if you feel like some of the goals are misguided or they should be changed or any part of this you know please queue up and speak up so that we can figure out if there's a way for us to be inclusive about those requirements so I'm going to basically. ✪
Manu Sporny: What's been driving the test Suite the VC API test Suite one that this groups working on to date and then you know we can maybe can compare contrast with traceability test suite and then and then go from there okay so. ✪
Manu Sporny: Link here we go ahead and put this in here in the chat channel one of the things that we're trying to do is just make it make it possible to test the VC API and the surrounding ecosystem in a modular way so basically we're trying to separate like the issuing test the things that just test the issuing end. ✪
Manu Sporny: Point in the options that you can set. ✪
Manu Sporny: Endpoint we're trying to test those in isolation right as one of the modular test Suites. ✪
Manu Sporny: Another modular test Suite would for example test of very specific crypto sweet so like the Json web signature stuff or the BBS plus stuff or ecdsa or ed2 55:19 signature 2020 those are examples of modular test Suites that just test one crypto sweet then we have modular feature test Suites so things like credential refresh. ✪
Manu Sporny: Artists list 2021 things that basically test a specific spec right a specific you know thing in the goal here or at least the vision with this set of test Suites is that if we get enough test Suites together we will be able to put together some sort of interactive interoperability dashboard so something that looks like can I use.com so if I go to like. ✪
Manu Sporny: This is a wrap up of just like thousands of tests that are run for a specific feature for browsers so this cannot use.com is something that web developers use to figure out if it's safe to use a particular feature or not will it be if I use this feature in my web page will it be supported across all the browsers so what we're trying to do is we're trying to build this kind of thing for the ecosystem driven off of the VCA. ✪
Manu Sporny: So what you can see here is Through Time different versions have different levels of support like for ecmascript 2015 yes 6 Chrome versions for through 20 didn't supported and then there was partial support from verses 21 through 50 and then finally at version 51 through 100 3 it got you know really good support with little notes like so it doesn't support tail call right this little to here that's probably you can't see. ✪
Manu Sporny: See but they're notes here for you know how things. ✪
Manu Sporny: These sorts of things help developers figure out what they can Implement in they also can help executive teams figure out what kind of product you know roadmap things they should they should contemplate you know has the as the years roll on Patrick go ahead. ✪
Patrick St-Louis: My question around for here we have a case you want to test an implementation against feature specification is there a concept about how to test an implementation against another implementation on the line let's say a company a has an issuer and Company B has holder company says a verifier are you just going to take for granted that. ✪
Patrick St-Louis: Towards the specific features they should be interpretable with one another or do you consider down the line having a sort of a flow driven test had added to these sweets with one implementation against another. ✪
Manu Sporny: That is an excellent question in I'm going to fast-forward really quickly and I haven't even looked at this part of the test Suite well no we don't do it there we do it somewhere yes the the short answer is yes there will be an N by n Matrix of one issuer calling another verifier to verify right in that's one of the things that the traceability folks are focusing on right now which is can you do workflows full full workflows so the short answer. ✪
Manu Sporny: For example the very simplest first workflow is something like did auth so did authentications defined in the in the BPA BPA R-Spec we will end up having a test Suite very shortly that does that did our thing and then the next thing is like okay well can you did off and then can they can the the issuer ask for a credential you deliver the credential and then they ask for another credential and you delivered that and then finally they give you the final verifiable. ✪
Manu Sporny: Credential so yes absolutely you know workflow. ✪
Manu Sporny: Exchange flows are contemplated in the future. ✪
Patrick St-Louis: Okay interesting are you familiar with the area Center probability information page. ✪
Patrick St-Louis: It can give a rough idea so it's a sort of Matrix towards all the different implementation so they respond different agents you know and then they test each agents against one another for given rfc's and at the end they give like a compatibility Matrix of different features and then if you go on the left you have each implementation on their own. ✪
Patrick St-Louis: Areas like there's the pet python. ✪
Patrick St-Louis: I work you know there's more popular ones and then you can dive into the details a lot deeper but I'm assuming is something a bit similar to that to you you want to do or. ✪
Manu Sporny: Maybe I don't know if it's someone with more of a familiarity with the Aries interop you know thing can look at it at a high level I'm looking at this and I'm not seeing features tested individually but maybe that's. ✪
Manu Sporny: Yeah so these are workflows being tested not necessarily features like jewelry first. ✪
Patrick St-Louis: Yeah this is very much one implementation playing a specific role in the interaction. ✪
Manu Sporny: Yeah can you're saying this is the interrupt Matrix like occupy against Occupy and if J against find e for example. ✪
Patrick St-Louis: Yeah and it's so they're based on the. ✪
Patrick St-Louis: One second here so if you have a look so the user as a reporter its Allure of reporter so when they run the test the use this other reporter in this submit the results at this endpoint there and if you click on the left you can find all these these projects and how they generate that Matrix page is the query the API of this sort of yeah yeah. ✪
Patrick St-Louis: So at the end of this sort of query so this sort of Allure of Docker service there's an API available that you can go query the results of every project and it's they have like sort of script that goes query certain and points and that's how they generate the interrupt page so the here it's a bit more in detail I see in your Suite. ✪
Patrick St-Louis: I'd like a custom reporter like. ✪
Patrick St-Louis: You see a proprietary reporter. ✪
Manu Sporny: Yes that's one of the ways it can be rendered and we'll get to that here in a little bit and looking look at how that's done. ✪
Manu Sporny: Okay but this is very very helpful so I think one of the differences I'm seeing here is that it's not clear which features are being tested and which parts of each sex being tested in which you know specific a IP is being supported other than the whole AIP right yep yeah yeah yeah so so this is good I mean we can we can look at this and compare contrast and figure out you know what's missing. ✪
Manu Sporny: Useful to have you know there as well. ✪
Manu Sporny: Okay back to the goals though that's I think this is kind of where we want to end up with the ability to drill down so some variation of the area's interrupt test and the can I use.com site and in and the ability to really drill down into very specific test to see where exactly you're failing and where you're passing and things of that nature. ✪
Patrick St-Louis: Sorry I have another question do you intend these test Suites to be available as tools for individual to test their only application where it's more to offer sort of represent Erie about current implementation and their status. ✪
Manu Sporny: Oh I don't know if I understand the question. ✪
Patrick St-Louis: Because right now like when I go into test Suite I can run the test suite and it's going to copy the VC API implementation repo and run the test against all those implementation but what if I'm an individual and I want to use this this sweet and preparation of some assessment was that the intended purpose of this test Suites okay. ✪
Manu Sporny: You can absolutely do that like we have people that download and run these tests by themselves on their own software or other people software one of the things we were trying to do is just not require people to have to do that this this these test Suites run on a weekly basis and we're probably going to ratchet them down to run on a nightly basis so you will be able to see whether or not certain implementers you know support. ✪
Manu Sporny: So you can use that you can use the test Suites and Target your own systems with them like if you have an internal system and you don't want to make it publicly known that you have an internal system you can use it to test the internal systems but what we're really trying to do is generate visibility for people that need to use these systems to see which vendors are compliant and which ones are not effectively. ✪
Patrick St-Louis: Okay so it's sort of like a to say French we say a deacon Technologic like a showcase for available in transition. ✪
Manu Sporny: Yeah that's right yep just like this is you know that for web browsers so you have all the web browser kind of vendors up at the top and then you can see you know how many of them support es6 for example right and they're you know they're thousands of features that are covered here like how many of them support web transport right we can see that web transport isn't doing that great right now except in Chrome and. ✪
Manu Sporny: And I'll put these two links into the chat fundamentally. ✪
Manu Sporny: What we trying to also do is create something that can be used in the working groups in like the verifiable credentials working group and in well yeah the in the VC working group we've got the 20 were coming up and it is a requirement w3c that you test every normative feature in the specification to exit to get to a global standard in so at some point we're going to have to write tests for the new. ✪
Manu Sporny: Data model the new data Integrity stuff. ✪
Manu Sporny: And that can be a very time-consuming and expensive process and so if we have something in place that allows us to test those things well before we need to go to Canada Trek or recommendation at w3c the better off we are right we can we can start demonstrating that these things that we're talking about in the working group are actually implemented multiple people support them the way you achieve that. ✪
Manu Sporny: And let me use an example here we see API it sure well I wonder if we have the VC test Suite in the past this was for verifiable credentials 10 we end up generating something that looks like this right so we take every single normative statement in the specification and we test it for a positive test and a negative test to make sure that people have actually implemented the specification so this. ✪
Manu Sporny: Patience so we test the basic document features and then we test the credential status features and then we test link data proves and credential scheme and so on and so forth right the problem with this of course is that we ran this test suite at the very end of the working group this is from 30th October 20 21 and we never really ran it after that was so so you don't actually know what people are how they do today unless you go and you run. ✪
Manu Sporny: One of the things with the design of the tests we did is we wanted to be able to run it on a nightly basis against systems that people were claiming our their latest you know greatest you know QA version at the at the at the very least Patrick please. ✪
Patrick St-Louis: Is this this sweet still relevant the VC data model test Suite or is it slowly moving away to the VC API. ✪
Manu Sporny: That's an excellent question I would argue in this is debatable that it's not relevant anymore because we haven't run it in a year and a half we could run it again but we would never be able to get the type of coverage we're starting to see with the VC API so I think you know what what what some of us are going to propose at the upcoming w3c TPAC meeting is that we use the VC API to run the tests so we. ✪
Manu Sporny: I will create a feature specific test suite for. ✪
Manu Sporny: Verifiable credential data model that will run these tests so we'll run the same tests but the mechanism will use to run the test is not asking each developer to run it on their machine generated Json file upload the Json file we will just basically say put up an endpoint that we can call on a nightly weekly basis and it will run all of these tests against the credential that you're issuing to make sure that you are a showing a valid verifiable. ✪
Manu Sporny: You can still run this and it should still work but because we're working going into verifiable credential 20 data models face it'll be it'll have to be revised you can still look so go ahead. ✪
Patrick St-Louis: It's a no it's just like I said the lab like we have clients coming in and you know they want to be not certified but they want to have like some sort of stamp than are compliant with this standard and that standard so we looked around what test Suites are available so let's say we have a client engagement and they want to be like sort of WT CE certified without going public just yet for example so the moment we sort of look at which test Suites are available and then we run the. ✪
Patrick St-Louis: As we give them a report we say yeah it looks good so. ✪
Patrick St-Louis: Test does it have some sort of validity in the market. ✪
Manu Sporny: When I say that I say that having been being the person that wrote a good chunk of the test Suite. ✪
Manu Sporny: Yeah so you know you can and you can get a gut reaction like if you have all kinds of red x's on your implementation there's a problem there but no I think we need an updated test Suite which is why we are suggesting let's let's use the verifiable credential API because it gives us repeatability and then the only thing that needs to be done is the just these tests need to be ported over which are is being done right now right so I'd say. ✪
Manu Sporny: About another four months we'll have something where you could run it against a certain infrastructure and get a good idea on whether or not their conformant. ✪
Manu Sporny: So it's coming if the working group agrees with the direction right yep and if they don't it's someone's going to have to propose something better okay so so we have this own the reason I went here is w3c expects these reports to be generated in a certain fashion but clearly this is not the only way that you can report how a test Suites. ✪
Manu Sporny: Reporter like this is another way right that Aries is providing this is yet another way of showing how the tests are running so they're different rendering things what we're trying to do is get the test Suites to get into a system where we can spit out data and then the data can be processed by test Suite result renders of which there are many so that's another part of this is this is one such rendering but it's not the only rendering. ✪
Manu Sporny: And should not be the only rendering the test Suite. ✪
Patrick St-Louis: Yeah if you showed a lot of page again and you just click on the sweets show all. ✪
Patrick St-Louis: Right there yeah so there you get a bit more details about specific so I think this gets more into. ✪
Manu Sporny: Yep yeah and then the difference here I think is that I don't know if these are actually tied back to the specification with w3c you're required to test the normative statements that's what tells you whether or not it's a it's an approved by the working group test or not do you know how that's here work or these directly from specifications or is this like someone decided that this is a good test. ✪
Patrick St-Louis: I don't mean to be a discussion in itself but basically the there's rfcs that go through life cycle so this is the page where they publish all the rfc's their request for comments and then in the Ares agent test our Nest there's different tags correlating to the rfc's themselves. ✪
Patrick St-Louis: The features or the concepts you can add specific ifc's and the test harness their specific tags that going to mention while if you want to run all the tests that are about accepted or if C or if it's a work in progress or depending on the maturity of the RFC you can test these specific test so it's a behavior driven test Suite so it works. ✪
Patrick St-Louis: Behave it's a python it's written in Python. ✪
Patrick St-Louis: And yeah it's a there's a lot to it. ✪
Patrick St-Louis: But I'd be happy to talk about this maybe another time. ✪
Manu Sporny: Okay okay yeah so but at the high level what we're trying to get to the design is designed around making sure that we've got something that's agreed by the working group officially testable on the normative statements in repeatable on a consistent basis with exports to different data formats that can be rendered in different ways fundamentally you know I think that's that's what we're trying to get to. ✪
Manu Sporny: Okay so that's you know about design the the there's a mocha w3c interop reporter you can generate other kind of plugins to render it in different ways and then we have this implementations directory which is where implementers come in and create pull requests for different infrastructure right and so if we look at like one of the most recent ones is measure I owe this is where they list their oath to. ✪
Manu Sporny: And of course we don't store the secrets and clear text those are protected secrets we have you know they're issuer and points that they allow in then the tags they can tag different ones with different things like this endpoint supports the Edwards curve 2018 signature the verifiers you know this says the tag say this is an oauth2 endpoint it supports the VC API. ✪
Manu Sporny: And then the company names measure. ✪
Manu Sporny: Easy so I think measure is our most recent integrate T it took a day and a half for them to integrate that tops and it's mostly because we had to like debug some of the oauth tokens that you know the oauth2 stuff wasn't happening appropriately so we had to look into why that was but it's meant to be fairly easy and once you list your implementation here it'll work off of all you know all of the different feature test Suites that. ✪
Manu Sporny: So that's design next item up or any questions on design or how we're going to get into individual test Suites now any questions on design before we get into individual test Suites. ✪
Manu Sporny: This is the latest run of the issuer test we'd so this this test Suite remember these are modular this only tests issuance that's the issuer API endpoint there so it's one end point and adjust tests the options that you send to that endpoint in that it actually generates something useful that comes back so the tests are really really simple here and as you can see we've got six one two three four five six. ✪
Manu Sporny: Implementations listed so far with another. ✪
Manu Sporny: You know a check mark means that they pass the test so for example this just texts check test to make sure that if you give it a valid credential the issue with valid options that it'll issue something and that's it that's all it all it does right and if you go over you hover over here you can see why certain implementations are failing right so at least it gives you data it may be inscrutable data but. ✪
Manu Sporny: At this implementer has made the credential subject a text string which is a violation of the specification so that's why this one's failing it's not an object it's texturing in in that's feedback right that's just one test and then there are other things that they pass like the request body must have a property of credential that's that's the request of the issuance and point the response must. ✪
Manu Sporny: That context defined that must be in a. ✪
Manu Sporny: That must include only strings that you know the credential must have a type property that must be an array and so on and so forth right and again this is just testing some basic stuff it's not testing for example credential reissuance it's not testing revocation it's not testing the digital signature that's on the credential itself and so on and so forth right. ✪
Patrick St-Louis: What about the other endpoint specifying and the DC API specification sources like their credential status and point which is meant to change the status of the credential is that planning to be included in this test later or is it limited to the credential issue and point. ✪
Manu Sporny: Yeah excellent question this is limited to the credential issue and point that's it. Right and so the the credential status in point will be a different modular test suite and these things are fairly easy to put together but that's that's a different test Suite it'll have different tests different normative statements in here that's tested so for example like this one is the verifier endpoint in the only thing this test is the verify and point in any options that you pass. ✪
Manu Sporny: Checks to make sure that invalid credentials will not verify right so there are all these tests that are done here it must not verify if the context is not an array must not verify if the items are not strings it must not verify if and so on and so forth right and then we've got a couple of failures here where the verification should have failed but it didn't for some reason right. ✪
Manu Sporny: So this is another endpoint credential status will be another endpoint the exchanges stuff will be another sorry another test Suite go ahead Patrick. ✪
Patrick St-Louis: I ran it a few times in the so the Mavs net there's ever that's gets reported and I had the same error on my test and it's a very strange error message so I don't know this is the right place to discuss that maybe open an issue but it says message to not exist and I'm not too sure. ✪
Manu Sporny: You may want to raise an issue on the test Suite in someone like this one Tachi might get back to you to talk through like to look at why it's failing right one of the other things that might be good is if you have an endpoint go ahead and submit your implementation so that we can start running it so that other developers can kind of try and debug why that you know what it's getting back from the end point and what it's expecting to get get back and so on and so forth. ✪
Manu Sporny: Of them already it's like this one's just inscrutable like who knows what's going on here right and so we're going to try and improve the error reporting here. ✪
Manu Sporny: Yeah I think that's that's you know we have to improve the airport in here right this is you can't you have no idea what's going on here. ✪
Manu Sporny: But fundamentally you know these are / in points modular tests that test the specific the specification itself right we also have other modular test Suites for testing things like digital signature so this is specifically a cryptography sweet test for the Ed to 55:19 Signature 2020 test Suite we've got you know it tests the issuer side of this it tests the verifier will hold on. ✪
Manu Sporny: On it tested at the data Integrity level. ✪
Manu Sporny: Meaning the data Integrity spec level and then it tested at the specific crypto Suite level so it tests to make sure that you know the byte ranges are proper in the valid you know a valid signature verifies has as valid and an invalid signature does not verify. ✪
Manu Sporny: Make sure that you're throwing certain errors like malformed errors if you have a mouth Forum signature same thing it checks to make sure that the proper data set canonicalization algorithm is run and here's the interop here's the N by n Matrix Patrick that you asked about before right so this is if one person issues a credential with this signature on it can the other systems verify. ✪
Patrick St-Louis: Yeah yes it does a lot so that's what the issuer name environment variable is for it basically decides which issuer is going to issue the credential to verify. ✪
Manu Sporny: Well it'll do an N by n so any it'll run all of the implementations against all the other implementations for this specific feature. ✪
Manu Sporny: So for the Edwards curve signature it'll have one and issue it and another and verify it. ✪
Manu Sporny: It'll do an end by in test against all of them. ✪
Manu Sporny: And I think the reason these folks are not passing it as they don't support the Edwards curve signature but for some reason they're included in the in the tests will have to look into that anyway this is basically to show that we are doing in by in testing as well here and then finally we're using the same infrastructure to do did method testing as well right and so this is very early days it has nothing to. ✪
Manu Sporny: To do with the VC API but is kind of a day. ✪
Manu Sporny: Early modular infrastructure that can be used across. ✪
Manu Sporny: Many different types of VC did you know ecosystem tests and they're more of these modular test Suites on the way these are just the ones that you know we feel comfortable sharing you know at this time there are ones that are we're working on for credential status so revocation suspension and they're also ones that were working on for credential refresh. ✪
Manu Sporny: They will of course be ones that are very specific to certain types of credentials and certain type of credential flows like maybe carry we might want to talk about one for the jff plugfest or some kind of Workforce credentialing flow or something like that the sky's the limit. ✪
Manu Sporny: That's that's primarily it any any other questions or concerns around kind of the test Suite it status kind of roadmap where we're going with it what we're trying to achieve. ✪
Patrick St-Louis: This is very interesting thank you. ✪
Manu Sporny: No problem thanks for all the good commentary and questions. ✪
Patrick St-Louis: What kind of help how's it like I do you are you looking for to help for this this is obviously being an open source program are you open to contributions. ✪
Manu Sporny: So absolutely yeah I mean we'd love help on anything in here so if anyone can see anything to improve here I think the biggest help we may need is creation of new test Suites for things that people want to test right so these are set up so that you should be able to copy the test Suite just the base template for it rip out all of the tests and put new tests in right so if you wanted to. ✪
Manu Sporny: To test a different type of crypto sweet like. ✪
Manu Sporny: Your signatures or you know SEC ecdsa signatures we need test Suites for that PC 20 data model will need test Suites you know for that so yeah Patrick if there's some some place that you'd like to help we have no shortage of work that needs to be done on these. ✪
Patrick St-Louis: Because one of our project we want to build this sort of interpretability bench you know that's sort of the generic term and so far like the Aries agent test harness and the test Suites of the liberal trees here the two most concrete sort of. ✪
Patrick St-Louis: Width and I know we are looking to hire a test that the mation developer in the we certainly might be interested in making some sort of contribution at some point. ✪
Manu Sporny: Yeah please do it would be it would be very welcome so you happy to talk more I mean if you want to put aside you know telecon time to talk about that or just you know work through things and get Hub issues we'd love to love to chat about that. ✪
Patrick St-Louis: Is there a channel for these type of conversation or these meetings are the only platform for this. ✪
Manu Sporny: So there's an IRC channel the GitHub issue tracker in the credentials community group mailing list or probably the best places to have a conversation. ✪
Manu Sporny: So for synchronous communication it's this these calls for asynchronous communication that's throw around ideas it's the credentials community group mailing list and then if for very specific targeted conversations GitHub issues or probably the best. ✪
Manu Sporny: With that and because we want to try and stop before the top of the hour I think that's the call for today any last words before we end the call. ✪
<kerri_lemoie> Appreciate this!
Manu Sporny: Okay thanks everyone we will get back to issue processing next week during next week's call and then the following week we will be skipping the call because of the w3c technical plenary in Vancouver okay thanks everyone have a great rest of the week take care bye. ✪