DISPATCH IETF-101 minutes March 19, 2018 9:30-11:00 London, UK Chairs: Cullen Jennings, Mary Barnes, Murray Kucherawy Thank you to John Levine for notes. Meetecho recording: https://play.conf.meetecho.com/Playout/?session=IETF101-DISPATCH-20180319-0930 JSON Content Rules. Andy Newton ( 20 min ) Martin Thompson: Is this done? AN: Very close to done. MT: Many other json schemas, why this one? AN: Want something specific to JSON, not over general. Neil Jenkins: This is just syntax, not semantics? AN: Yes, have some semantic proposals, want to solve this first. Phill H-B: I've done some of these, this spec is too complex, e.g., don't need to limit integer range. EKR: we've had a lot of these, do we need one more? Who will use this? Needs to be more than 1. AN: People have come up to me and expressed interest. Colin Perkins: Define so you can extract from RFCs AN: Yes, doing that. Fluffy: who would use this? AN: routing info in RDAP Richard Barnes: acme needed something like this Pete Resnick: why is this different from json-schema? AN: different approach, json-schema is pure json, this isn't to make it more readable MT: maybe for jmap, still concerned about NIH and too many of these Adam Roach: some CBOR work generating JSON NJ: jmap has descriptions, haven't found spec language useful, just need types Mark Nottingham: have use cases, need something like this Hum for any sort of json schema? yes  Joe Hildebrand: CDDL in CBOR WG is similar to this EKR: don't want multiple competing specs PHB: don't need to tie schema of serialization JH: could take CDDL out of CBOR WG Mary Barnes: take this to the list Defining Well-Known URI. Mark Nottingham ( 20 min ) PHB: .wellknown is like a port number Ben Schwartz: DOH WG is thinking about discovery Sam Weiler: could be a less bad better option Ted Hardie: gatekeeping via registrations for schemas have tried to keep stuff clean, people ended up ignoring it and colliding, changed the rules consolidate this into BCP190bis? MN: much sympathy for open registries in most cases TH: where it's an interop issue, gatekeeping makes sense, which is this? Cullen: what is common in address spaces we've rejected? is there a common question? Paul Hoffman via Cullen: need time to discuss this Carsten Bormann via Cullen: do we need a discovery space MN: that's host-meta Dan K Gillmor: outside the IETF nobody knows what a URI is.  User interfaces using URIs don't work any more Bron Gondwana: email address domains give us a hostname and port, we don't want another discovery system PHB: want to be able to connect to a web service using DNS discovery, can do it with SRV, same prefix as .well-known Jim: we deprecated WKS a long time ago Roy: question was whether SRV names should be directly under .well-known or under an intermediate name MSK: we have work to do, needs mini-WG          TH: consider both this and bcp190-bis Cullen: yes Andrei: mail the ART AD's, they'll concoct a charter           Rethinking a new media stack. Cullen Jennings (20 min) Roni Even: good to have an overview of the whole problem can we converge RTP and HTTP streaming? Colin Perkins: network when we designed RTP is not our network now Harald: disagree with approach, need compostable components CJ: do these approaches lead to the same place? HA: components will be deployed faster Jabber John Klensin: how will these changes get deployed? They're significant changes CJ: let's set up a discussion venue Mary: new mailing list Adam Roach: Cluster 238 of dependencies nearly done Opportunistic message and email encryption. Bernie Hoeneisen ( 5 min )           - draft-birk-pep-01           - draft-birk-pep-trustwords-00 This is mentioned here but discution should be in security area.