Minutes - GEOPRIV - IETF 85 Summary 1. WGLC for draft-geopriv-flow-identity WGLC completed with several expressions of support from the room. 2. draft-jones-geopriv-sigpos-survey Continued interest in this draft from the room. Kipp will rev before we issue call for adoption. 3. self-published-geo Development of yet-to-be-published draft will continue, alignment with geopriv formats possible. Raw Notes from Roger Marshall ----------------------------- Geopriv Notes Chairs: Alissa Cooper, Richard Barnes Wed, 14:40-15:40 Notewell Covered Agenda Bashed Document Status (Alissa) geopriv-flow-identity – need at least one more person to say that it should go to WGLC Kip Jones: concur that it is ready to go Brian Rosen: also have read, concur it’s ready for WGLC Alissa: Ok, thanks. --Presentations: -- draft-jones-geopriv-sigpos-survey Kip Jones presented Henning Schulzrinne: 2 kinds of signal info – Access Points & Metrics, e.g., Android phone measures – are you talking about both? Kip: yes, both. More general case is the survey – more focused on a professional survey, not so much for the crowd-sourced model. Henning: One of the big concerns was how do you trace the information back. Kip: more explanation scenario Henning: Metadata - Want to be sure that data is properly attributed, also want to know that the data is properly represented, on a known schedule, update frequency, expiration of the data, notion of trust of data vs. “use extreme caution”. Martin Thomson: question the validity of data – “expiration” may be the thing you don’t want. Henning: need a way to characterize – need to know what “expiration” means. AP may/may not be where it was 3 years ago. Martin: reason why expiration is important to define Bernard Aboba: Rights to the data covered? Martin: Licensing gets interesting, essentially talking about copyright. Kip: Licensing is an important piece, but also a fuzzy area currently. Metadata is important and will need to be looked at. Kip: Henning: well imagine that Universities may not have objection to Creative Commons, but want to be attributed. Kip: Looking at Geo DRM – pretty heavy right now, may not need to be as heavy – need to look at it. Alissa: Not just licenses, right? Kip: Looks at enforcement – not a requirement now, but points to the ability to do this kind of stuff. Kip: Derivative rights, thoughts about location data being unencumbered. Martin: There is a thought of the location being restricted to rights. Impacts to the Rulemaker. Kip: Access control part of this – idea that the location may be unencumbered is the user has authorization to get it. Kip: Happy to get comments. Henning: You may want to look at Open Data Commons Attribute licence Kip: Yes, did take a look, will need to look at it further. Kip: Went over changes in version -01 Kip: Known Issues (slide) Kip: Additional issues (from Brian Hart), includes: Definition of a LIS still needs some work “_nomap” for AP privacy – don’t know how you would put this into the document Henning: questions on this Richard: This is something – a consideration that would go into the privacy section. Kip: Also, to Henning’s point, not sure where the ownership belongs James: Suggest a fairly lean privacy section, and then maybe use a separate document to go deeper. Richard: I don’t think we’re going to list all the privacy laws in this document. Henning: I see this as more of, “we are aware that these things exist”, etc. do your due diligence. Even in post cleanup – if you knew of it at the time, you shouldn’t have gathered it, and marking it isn’t going to do anything to help. Kip: binary version Brian: Please don’t do that. Martin: we have this gzip – and it produces binary really well. Kip: Okay. Now with Brad engaged, I’ve got a cohort to work through these issues. And the licensing issues need worked through. Alissa: Are there other non-IETF folks out there that are willing to help? Kip: Yeah, I think we can engage more – but a chicken and egg thing. Richard: It’d be good to get some of the companies doing this surveying Henning: Yeah, Kip’s company is one of those Henning: It seems at least two options, one: punt, we’ve got tools for, go at it – you don’t want to ship data files every night, or two: publish/subscribe model Kip: How do you propagate this document to all the interested parties. Henning: Seems like a much harder generic problem – probably have business’ that are data aggregators – seems like well beyond the scope of this… Henning: Speaking from Public Safety model, down the road – far-fetched at the moment, the Fire Marshal, next to checking fire suppression, could be checking floor plan data, etc. – like I said, far-fetched for now. Richard: Seems to be some interest in the room. --next Presentation Self-published-geo Eric Klein presenter (not a regular geopriv follower, usually do IPv6) IP-Geo – people have different views, but at Google people get really pissed if it routes the user incorrectly. Eric: example scenarios given Examples: https://registration.icann.org/geo/goole.csv 199.91.192.0/21,US,US-CA,Los Angeles, 2620:f:8000::/48,US,US-CA,Los Angeles, Martin Thomson: Thank you – explains just how _unsuccessful_ this work group has been! Eric: This is for special use. Martin: When I travel I send a header that specifies EN, English – and get local context based on IP-Geo Eric: May be useful if there are local data constraints Martin: A lot of sites Richard: This is one input, a channel to push published geolocation. In terms of geopriv context, when compared to geopriv’s data elements, aligns quite well. Richard: how do you find the URIs? Eric: 3 of them, Ad-hoc, Various “$-sign” databases (e.g., Whois), Reverse DNS, etc. Richard: which mechanism is recommended for a user? Eric: Henning: Distributed databases mentioned, privacy mentioned, DSLAM level information, restricted for Public Safety use only, and we want to figure out how we provision the data. Also, rights, sources, metadata – across the board issues, also, who wants the data – may be good to bring efforts together, rather than treat as two separate universes. Andrew: Surprisingly good to located – at least in the UK example. Eric: Depends on the region Alissa: Thanks for presenting (End of Meeting) 15:40