<!DOCTYPE html>
<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=UTF-8">
</head>
<body>
<p>Here are the notes I took yesterday. Please check that I listed
all the attendees.<br>
</p>
<p>Attendees<br>
Manolis Viennas<br>
David Chadwick (notes)<br>
Torsten Lodderstadt<br>
Kristina Yasuda (Chair)<br>
Judith Kahrer<br>
Philipp Lehwalder<br>
Nemaja Patrnogic<br>
Jan Vereecken<br>
Brian Campbell<br>
Micheal Jones<br>
Oliver Terbu<br>
Pedro Felix<br>
lukasz Jaromin<br>
Juba Saadi<br>
Javier Ruiz<br>
John Bradley<br>
Paul Bastian<br>
George Fletcher<br>
Daniel Fett<br>
Tom Jones<br>
Daniel Fett<br>
Ben Piercey<br>
Brian Campbell<br>
Sébastien Bahloul <br>
</p>
<p>Manolis introduced himself, from Scytales.</p>
<p>VCI PR#276 (Define claims display description and claims path
query)<br>
Daniel explained that this changes claim names into an object with
a path and claim name. This caters for both nested claims and
arrays of claims. An addition to support object types has also
been requested. It should work for both X.509 and CBOR encoded
claims. There is already an example for an ISO mdoc. Kristina said
that using null is fine, but should the array have mixed integers
and strings for values? Joseph said the alternative is to use
strings everywhere, but this wont help to differentiate between a
real integer and a real string. Pedro thinks using the same model
for authz details and metadata might introduce difficulties for
odd/edge cases. Also wouldn't it be better to have different
object models for Authz details and metadata description that
share the same path description rather than putting exclusions in
one of the object models? Joseph said this should be an editorial
change. John asked how does the issuer specify a policy for the
wallet that constrains what it can display to untrusted verifiers?
This does not seem to have been discussed anywhere. Torsten said
that John's issue is about the issuer controlling what the wallet
can display/transfer to verifiers, but this PR is about how the
wallet displays to the user details of the issued credential. Paul
said the authz details description should not be linked to claims
description but only to claims path query. Kristina said this is
an editorial issue.<br>
</p>
<p>Kristina has requested reviews on this PR from a large number of
the WG attendees.<br>
<br>
VP Issue #112 (Opportunities to improve the credential
query/request syntax)<br>
Kristina introduced this issue and asked people to state why they
do not like DIF PE.<br>
Should we keep PEv2 or work on PEv3 or define our own query syntax
or just have this as an extension point for any query language to
be plugged in.<br>
Mike stated that its complexity e.g. algebra can lead to interop
problems.<br>
Torsten stated that the VP protocol supports credentials in
multiple formats, and PE allows the format to be stated first
before the query is presented. Many people have implemented PE so
dropping it can have large consequences.<br>
John stated that a unified format independent query language is a
strength of PE.<br>
Brian supports the extensible model but Torsten says we have
failed in interop if we make it an extension point.<br>
<br>
</p>
</body>
</html>