20:58:11 RRSAgent has joined #auto 20:58:11 logging to http://www.w3.org/2017/02/06-auto-irc 20:58:13 RRSAgent, make logs public 20:58:16 Zakim, this will be 20:58:16 I don't understand 'this will be', trackbot 20:58:16 Meeting: Automotive Working Group Teleconference 20:58:16 Date: 06 February 2017 21:00:20 AdamC has joined #auto 21:00:50 PatrickLue has joined #auto 21:01:36 PatrickB has joined #auto 21:02:35 Present+ Adam, Paul, Hira, Ted, PatrickL, PatrickB 21:03:40 PatrickLue has changed the topic to: auto auto webex for feb. 6: https://mit.webex.com/mit/j.php?MTID=mb30566c23e018cb0ca9841c8fe263c30 21:03:45 scribenick: ted 21:04:55 KevG has joined #auto 21:05:32 Present+ Kevin 21:07:14 agenda: https://lists.w3.org/Archives/Member/internal-autowebplatform/2017Feb/0000.html 21:08:07 Paul: trying to organize these calls more 21:08:17 … given some limitations on screen sharing from Linux we resolved to try to have materials shared in advance 21:08:35 … Patrick or Patrick will present a comparison 21:10:03 -> https://lists.w3.org/Archives/Member/internal-autowebplatform/2017Feb/0003.html Comparison exercise 21:10:54 PatrickL: some types of developers aren't keen at looking at documentation, in automotive they generally are but we'll be attractive other types 21:11:35 … in viwi you can get all available with a GET of /car or GET of /car/door 21:11:59 … going the exploration instead of documenation approach you'll need to do two calls to find the doors 21:12:22 … to get a list of all open doors you can ask GET /car/doors/?isDoorOpen=true 21:12:46 … in viss to get a list of all doors that are open it seems you have to get all door data and iterate over them 21:13:25 … assuming you could close a door through these interfaces you would do a POST in both examples (set for viss) 21:13:38 [[POST /car/doors/ 21:13:38 { isDoorOpen: false }]] 21:13:48 in viwi 21:13:50 [[SET doors..isLocked: true]] 21:13:52 in viss 21:14:14 PatrickL: I understand filtering for subscriptions but not for get 21:15:17 (in viss) 21:16:06 Kevin: @@k 21:16:23 wonsuk has joined #auto 21:17:07 PatrickB: filtering can be very handy for media for example (when you have extensive audio files available) 21:17:54 Kevin: is this an action to add filtering on viss get? 21:18:08 Paul: this isn't a WG call, no changes are being asked 21:19:13 PatrickL: we might find things that are desireable to implement as a result of these discussions, we're just reviewing differences 21:20:31 paul has joined #auto 21:20:55 Paul: it might be best to collect these on a wiki 21:21:10 https://www.w3.org/community/autowebplatform/wiki/ViWi 21:22:36 [ted will rename the wiki to the name Rudi came up with 2G-VIS] 21:23:15 Ted: people are welcome to create additional pages or use this one and i'll likely rename this one and redirect it 21:24:23 PatrickL: Kevin, do you feel this sort of comparision is helpful or do you have a different suggestion? 21:24:59 Kevin: it makes sense. my concern is we want to avoid going back to square one with AGL or PSA... 21:25:29 Paul: Justin Park did a comparison of 5-6 different specs, focused mostly on signal level and find the core set of things that are useful 21:25:34 q+ 21:25:53 … you might want to take high level differences and different paradigms and bindings 21:26:07 … cataloging helps the discussion 21:26:49 Kevin: viwi handles additional things like we've been discussing like media 21:33:56 … I may shameless steal the idea of filtering on get for viss and adding that as an issue 21:34:48 PatrickL: we can do additional comparisons, wonder how helpful they will be 21:36:04 Consensus item - Consider adding filtering to get (as implemented in VIWI) - added as issue 135 21:36:17 rrsagent, draft minutes 21:36:17 I have made the request to generate http://www.w3.org/2017/02/06-auto-minutes.html kaz 21:36:45 regrets+ Kaz 21:40:20 Consensus item> Consider adding filtering to Set e.g. to set a subset of doors so that isLocked=true #136 21:40:26 Ted: @@@on_rest 21:41:09 Kevin: are you doing filtering on set (POST) in viwi? 21:41:15 PatrickB: it is possible 21:42:03 … the idea was to have http querystring that would restrict results as if it were a sql select 21:42:57 … you can similarly have a POST that sends a lock request on all unlocked doors instead of all doors based on querystring 21:43:22 Kevin: that is rather neat 21:43:47 PatrickB: you can similarly subscribe based on query parameters 21:45:53 Ted: I'm wondering if you can filter on attribute level on viss, think you can 21:46:28 Adam: yes, you can identify an attribute value you're interested in and subscribe to them 21:48:08 https://github.com/w3c/automotive/issues/136 21:48:13 https://github.com/w3c/automotive/issues/135 21:49:46 [comparison on screen share of trying to do a selective set/post based on attributes] 21:50:51 PatrickB: you can easily use both notations, CSS of VISS, or REST of ViWi 21:52:45 … we should bear in mind others' approach outside of automotive since we will be bringing them together in the same space 21:53:36 Kevin: do you see viwi wrapping around audibles, pandora etc? data will be downloaded from those services but then they would be available to app developer on head unit? 21:54:13 PatrickB: yes and no, some you won't wrap and just have eg spotify available as an api in a similar manner 21:54:59 … depends on the service you want to address. some are aggregrated from multiple sources 21:55:35 … viwi api can provide links to songs you want to play back that are in your library but you can't get a link to spotify song 21:56:07 PatrickL: disclaimer Audi is doing things different than rest of VW group at present 21:58:04 … assuming content providers are subscribed to and available on a head unit, you'll have a way to get to them from viwi 21:58:38 … I started to compile a list of possible comparisons in wiki 22:00:02 Paul: speaking from OpenCar point of view, I have seen various OEM SDKs and appreciate the abstraction layer and not have to integrate on a lower level 22:00:56 … we are not talking one size fits all here 22:01:16 … it can be helpful to see how sockets compare to rest to APIs/SDKs 22:01:45 … I'm willing to put together some notes from my experiences on the wiki 22:01:56 … people will come at this from different perspectives 22:02:23 PatrickL: it would help me to understand why this approach went sockets only 22:03:35 Paul: perhaps Adam or Kevin can help put something together for that. it was a group decision earlier 22:04:03 Adam: I think it boils down to we saw we had enough of what we need from sockets alone 22:04:14 PatrickB: I'll prepare some arguments for REST 22:06:35 http://doodle.com/poll/4fwebn3qdu39hpsu#table 22:07:31 Every Monday 2100 PST / 0600 CET / 1400 JST 22:08:58 [discussion on alternate time that is more friendly to Asia] 22:10:58 s/@@@on_rest/the biggest difference is sockets vs http rest+sockets. we dropped http before in interest of time and because we felt we had what we needed from sockets. that is worth revisiting/ 22:12:03 I have made the request to generate http://www.w3.org/2017/02/06-auto-minutes.html ted 22:13:35 https://www.genivi.org/events 22:14:02 s/that is worth revisiting/that is probably worth looking into further/ 22:14:03 I have made the request to generate http://www.w3.org/2017/02/06-auto-minutes.html ted 22:14:19 https://www.automotivelinux.org/news/events 22:14:24 I have made the request to generate http://www.w3.org/2017/02/06-auto-minutes.html ted