W3C

Shifting Into High Gear on the Web -
W3C Workshop
Day 2 - 15 Nov 2012

See also: IRC log

Attendees

Present
Adam_Abramski, Alejandro_Piñeiro_Iglesias, Alex_Ajao, Amy_Leeland, Andy_Gryc, Boris_Geller, Burkhard_Mayer, Chris, Christian_Müller, Daniel_Austin, Dave_Raggett, Diana_Cheng, Erik_Pellemeier, Haruhiko_Kondo, Hiroshi_Ito, HyungJin_Jeon, J_Alan_Bird, Jean-Marc_Temmos, Josh_Soref, Justin_(JongSeon)_Park, Kazuhiro_Kitagawa, Kazuyuki_Ashimura, Ludovic_Alidra, Magnus_Gunnarsson, Magnus_Olsson, Marcin_Hanclik, Marie-Claire_Forgue, Marius_Spika, Markus_Muenkler, Matthias_Bezold, Maximilian_Michel, Michael_Feld, Mikko_Ylinen, Mitsunori_Maru, Naomi_Yoshizawa, Nils_Oppermann, Oliver_Novakovic, Pavel_Konopelko, Peter_Gaus, Petter_Graff, Philipp_Hoschka, Prashanth_Halady, Raymond_Hoogendoorn, Roberto_Zompi, Roger_C_Lanctot, Roman_Steiger, Ryuji_Wakikawa, Scott_Pennock, Sebastian_Speiser, Shigeyuki_Sakazawa, Shinjiro_Urata, Simon_Isenberg, Tatsuaki_Takahashi, Tatsuhiko_Hirabayashi, Thomas_Rottach, Tsuguo_Nobe, William_Greenly, Woochul_Jung, Youngwoo_Choi, [CUT], _Delaney, Cheree_Anne_Schepp, Paul_Slattery
Regrets
Chair
Dave & Adam
Scribe
Bernard, Alan, timeless

Contents


Strategy Analytics - Car trends/forecasts

<Alan> Adam: Introduction of Roger Lanctot (Strategy Analytics)

Roger: Strategy Analytics intro
... multi platform environments, domains
... user experience expertise
... labs to do research
... very practical activities

[The Big Picture]

Roger: the big picture
... enhance the driving experience

[the car is a "persistent inquiry]

[Beyond the icon]

Roger: experience with Android car last week in California
... we think we already passed that Icon interface
... what's next, next experience
... the details

[infotainment design trends]

Roger: we are in gen 4
... hybrid solution

[MyFord Touch; 4th gen segment leader?]

Roger: Ford solution to allow user's phone usage
... sync solution
... hybrid solution connected phone and on board systems

[Tier 1 suppliers]

Roger: Software/services companies growing in this market

[the automotive cloud has arrived]

Roger: 3G, LTE is on the way, we have all the technology on the table to develop consumer experience
... social networking access is also there
... the car connnectivity in the car will be monetized
... connectivity will provide more services to customers
... (anticipating pbs is the key - not selling the apps)
... voice in the car: early phase for now
... demo experience on voice environment
... my message is "please think outside the box"

[HMI trends]

Roger: HMI: touch screens, sensors, language voice
... natural language to connect services

[connecting the car: the details]

Roger: Displays in the car
... driver distraction
... some discussion in US about displays
... 3G/LTE for CRM.VRM on the way
... iPhone integration with BMW/Mini

[Connectivity ECU-OEM]

Roger: connectivity ECU: forecast: 100$ in 2011, $75 in 2019
... LTE forecast
... Telematics ECU
... OEM display deployment

[OEM Display Deployment]

Roger: everything will be merged in the car

[OEM Headunit]

Roger: enhance the audio experience
... Harman is focusing on the audio experience
... from CD to high quality audio experience
... Digital radio is coming

[OEM Digital Radio]

Roger: Radio and some info services part of Nissan services
... Sirius experience

[Must have Infotainment-Western Europe]

Roger: Smarphone presence in the car

[Smarphone's impact on the in-Vehicle Experience]

Roger: BMW experience
... safety is the key

[Interest in Connected Services: US & Western Europe]

Roger: Facebook and Twitter: main usage is access to the data

[Mobile Phone Usage]

[Car Connectivity a Mobile Mess]

Roger: Nobody is doing right
... don't see what is in the market today

[MyFord Touch-Coming to China]

[GM CUE]

Roger: UI comments, usability of the icons ?

<dsr> icons are a step backward

Roger: and Now the videos !

[ RLink Video ]

[R-link Renault video]

Roger: no access to the CAN bus
... you have to purchase the apps, but the service is free
... Renault Zoe apps video
... RLink, you are not able to connect your phone to the system
... you don't want an app system you can only use in the car
... you need to be able to manage in the car and out the car usage, don't lock the environment
... Like Renault, Chrysler also wants to monetize via app store
... [Chrysler video]

[Video: Uconnect]

scribe: 4 solutions, 3 of which uses the phone for connectivity

[The future of HMI-Too Convential]

[Touch Screens & Voices Forecast]

Roger: voice recognition evolution

[Car Makers Need]

Roger: in the car you are a persistent query
... s/query/inquiry/

Q&A session

Roger: LTE usage

dsr: you said icons are the past, what do you think is the future?

Roger: services that run in the background anticipating and adapting to your needs.

Roger: Some works in progress
... software is the solution to the different market needs

<kaz> kaz: what kind of mechanism is used for modality integration? .. proprietary ones?

<kaz> roger: nuance should be proprietary

Roger: Android in the car as a container
... CloudCar solution
... regarding Apps, at the end need to control the content

<timeless> scribenick: timeless

Adam: thank you Roger

[ Applause ]

Adam: without delaying further
... our next presenter is from Vodafone
... can we bring up diana's slides from vodafone

Is it time to power Infotainment and Car Portal Applications with HTML5 and Device APIs?

diana: Good morning
... i'm a little ill, so no questions

[ Diana Cheng - R&D, Web Expert ]

diana: i'm a member of the DAP WG
... and recently joined SysApps as well
... i'm here w/ markus muenkler

[ Contents ]

diana: Introduction: Status of the automotive app ecosystem
... Web technologies as an enabler for Automotive
... Readiness of Web technologies: What is missing?
... Deep Dive into the enabling HTML5 features
... Conclusions: Key takeaways

[ The automotive technology ecosystem ]

diana: it's fragmented
... we start to see BlackBerry Tablets in cars

[ The Web Technology Ecosystem ]

diana: I'm using HTML5 broadly
... a lot of web technologies aren't part of HTML5
... I'm using it as an umbrella term
... We use it
... and we have to do some adaptation for different targets
... but it's still easier
... With a web application, you can immediately push updates
... For Web Applications
... mobile browsers support HTML5 to some extent
... they're supposed to support something, but it isn't that good

[ Web - So what's new? ]

diana: new web platforms
... starting to integrate more w/ the underlying platforms
... Tizen / Firefox OS
... more access to underlying data/hardware
... the native gap is closing

[ Are HTML5 and Browser-based APIs ready? ]

diana: we have the question is HTML5 ready?
... it depends on the UC
... for the FT
... if you see it on the iPad or tablet, it's a really good experience
... mobile web app from Facebook is still a really good app
... it's close enough to a native feel
... what do we need for Automotive?
... do we need all the power of native?
... probably not, most don't need it
... what needs to improve in HTML5?

[ Automotive Requirements: 2 Main Fronts ]

diana: there are a number of areas
... automotive can include user at home
... wanting to send driving directions from PC
... it's easier to search from home
... so some sync needs to be there
... speech recognition with offline support
... cellular connectivity can't be guaranteed
... Text-to-speech
... incoming notifications/traffic notifications
... Navigation/Routing
... if we want apps in the car, we want them to be installable and easily upgradable
... [Those were in-vehicle ]
... [Mobile Car Portals ]
... If there's car sharing in the family
... the main driver could receive notifications if someone else is using the car
... and there's an accident
... and access to the camera
... take a photo of a license plate and upload it to the web

[ IVI Applications - Session Handover (1/2) ]

diana: you leave your vehicle
... and you want to continue listening to the music
... or you park your car in a spot, and you need to walk to the destination
... your walking directions could be synced to your phone

[ IVI Applications - Session Handover (2/2) ]

diana: we need notifications to devices
... native platforms have Android C2DM/GCM
... Apple has APNS
... BlackBerry Push
... they all have restrictions
... Android has Message Quota
... iOS has no ads allowed
... -- Web Push
... We have a way in HTML5
... Server Sent Events
... - 1 way push, push standard dom events, no specific implementation needed
... integrated in the browser
... WebSockets
... - bidirectional channel between a server and a web app
... good for sending / receiving updates from the server as well
... if you leave the vehicle and your phone receives a notification
... the user needs to start the application manually
... there's limited support for these technologies in mobile browsers
... - I had to use a polyfill to see if WebSockets are supported and if not, fall back to XHR

[ IVI Applications - Speech Recognition ]

diana: we need speech recognition
... a lot of solutions offer `good` capabilities
... HTML 5 Speech Input
... there was a spec from the Speech XG
... there was a proposal to do best effort recognition, just as GeoLocation does

<inserted> [ actually, it is a CG ]

diana: there was a proposal for a new W3C WG for it
... there was minimal adoption
... only webkit, and it requires sending it to network service
... breaks the offline support requirement

[ IVI Applications - Text-to-Speech ]

diana: we need text-to-speech as well
... you need realtime directions
... there are apis you can embed in your web apps, e.g. Google Translate
... there's no standard
... and those require network connections
... some proposals...

[ IVI Applications - Navigation (1/2) ]

diana: a lot of sophisticated devices
... tom-toms, etc.
... HTML5 covers many of the requirements
... GPS via GeoLocation
... watchPosition() so you get updates
... WebWorkers for route calculations
... - but requires more mobile support
... Offline requirement?
... - having navigation without connectivity

[ IVI Applications - Navigation (2/2) ]

diana: how do we create offline applications
... how do we access maps previously loaded?
... AppCache - something that could help us
... - `widely documented to be a bit rubbish`
... if one element that you tried to get fails to download, you revert to the previous version
... WebStorage: Small quota, storage of binary assets is not easy
... has performance issues
... because it's a synchronous API
... Other APIs exist
... but they're making slow progress
... IndexedDB
... has desktop implementation, but no mobile implementations
... WebSQL was implemented in mobile browsers
... but i haven't seen information if new mobile browsers support it
... there's FileSystem/File Writer
... from Google, but Mozilla doesn't like it
... there's no locking
... Mozilla created their own File System API
... which extends IndexedDB
... there's fragmentation
... and no acceptance of existing proposals

[ IVI Applications - Installable Applications ]

diana: W3C Widgets
... no mechanism to automatically update
... lose advantage of updates re hosted apps
... w3c widgets aren't popular
... Google has their own, Mozilla has their own
... fragmentation needs to be addressed
... need to get people to converge
... PhoneGap as an alternative
... (now Apache Cordova)
... you can push updates to native stores
... it's a workaround, it's an open source project
... not a standard

[ Car Portals ]

diana: UC Insurance
... we have a dongle that goes in a car
... right now it's mostly static data
... it has a Global SIM
... so the user can move freely throughout Europe
... it has GPS, Accelerometer
... to log trips of user
... log driving behavior
... see if there's harsh breaking
... The Mobile Web interface looks like this

[ --Picture-- ]

[ Car Portals ]

diana: you can locate your car
... see your current position
... you can zoom in
... get directions

[ Car Portals ]

diana: receive error messages
... from the dashboard
... we calculate a Green score, and an insurance driving score

[ Car Portals ]

diana: i have the application, i can show it
... it behaves in modern mobile browsers
... Chrome for Android
... it behaves very nicely
... not so much so for Android 2.3 browser
... - issues w/ performance

[ Car Portals - Map Visualization ]

diana: we embed maps
... there are lots of apis
... Nokia Maps, Google Maps
... you can make rich UIs
... Touch requirements for mobile
... it doesn't work very well in slightly stale browsers
... Android 2.3 is poor
... no pinch to zoom, double tap- nothing [useful] happens - the page jumps
... `overflow:auto` problem
... you can't scroll the content in the middle if you have fixed element content
... Android 2.3/older iOS Safari
... you have this problem
... you can find polyfill with CSS Transformations
... but it hurts performance
... it's an implementation issue, not a standards issue
... not sure about the story of pushing updates to older platforms
... going forward, we want to avoid these issues

[ Car Portals - Real-time Notifications ]

diana: we want to send notifications
... if there's a crash/heart break
... native apps have status bar which grabs user attention
... important for emergency alerts
... we don't have them in mobile
... there's web notifications

Josh_Soref: BlackBerry PlayBook has it

diana: most mobile browsers don't support it
... not good for getting user's attention
... Vibration API
... to couple w/ Real-time notifications
... -it's coming to mobile browsers

[ Key Take-aways (1/3) ]

diana: a lot is covered
... we need to work to fix some
... for enterprise... UVI UC
... this technology enables `fast to market`
... and target a wide range of users

[ Key Take-aways (2/3) ]

[ Key Take-aways (3/3) ]

diana: need to agree on storage mechanism for large data
... WebSQL on mobile?
... IndexedDB?
... FileSystem API?
... what about building something on IndexedDB instead?
... it seems to have the underlying performance
... browsers need to support it
... Broader support for WebSockets - mobile browsers aren't there yet
... fragmentation in Web Packaging
... either changes to W3C Packaging,
... or get convergence from Browser vendors
... Need to standardize something for Text-to-speech
... with best effort and fallback to embedded solution
... best performance if connectivity is available
... Similar for Speech-recognition

[ Questions? ]

diana: that's it

[ Applause ]

ph: very nice talk about what's there
... and where the gaps are
... we see a lot of stuff is available/can be used
... things need to be pushed further
... The dongle you showed?
... is that something only you can use?
... or can others use the data?

diana: atm, we have a company that gathers the data in the backend
... and making it friendly
... and there's a REST api
... but it isn't so open yet, no.

Adam: thank you diana

[ Applause ]

KDDDI's position for Web and Automotive issues

sakazawa: Shigeyuki Sakazawa, KDDI corporation

[ Agenda ]

sakazawa: introduction of KDDI
... seamless service concept
... connecting vehicle and smart phone

[ About KDDI ]

sakazawa: KDDI is a japanese telecom carrier
... mobile telecom, fiber-to-the-home, tele com solutions
... content services and banking

[ KDDI's Convergence Strategy ]

sakazawa: multi-use
... WiMAX, 4G, cable-to-home
... Vehicle issues
... application + content services

[ Seamless service concept ]

sakazawa: we'd like to realize seamless service
... service anytime, anywhere, any-device
... service on a personalized basis
... service with more pleasure
... during an event: watching video: Passenger, shopping, driving
... - including before/after event
... For driving-event
... Make driving plan and download maps (@home)
... Maps/music/other data downloaded in your smartphone (@driving)
... search nearby (@cafe)
... restaurants, shops, other data (@driving)
... - driving is part of our life
... want to connect everyday life to event

[ Requirements for seamless service ]

sakazawa: contents & Apps ubiquitous use
... vehicle personalization
... seamless networking

[ Functions to be realized ]

sakazawa: contents and apps ubiquitous use
... - speaker out of music from smartphone
... show a navigation map stored from smartphone
... recall address book from the phone and display on headunit
... - to make a phone call
... Vehicle personalization
... personalize various settings for a rental-car
... and erase personal information from shared car
... tune User Interface
... not just menu appearance, but also speech recognition

[ Functions to be realized ]

sakazawa: seamless networking
... continue streaming and sequential update
... prefetch heavy data before

[ Characteristics of headunit/phone ]

sakazawa: all-in-one head unit
... smartphone
... hybrid smartphone + head-unit
... + low cost, up to date service, high usability
... - high standardization required

[ two approaches for hybrid ]

sakazawa: Display Image Transfer - MirrorLink
... distributed processing
... apps run on headunit
... headunit+phone share processing

[ KDDI's prototype - Web Infotainment ]

sakazawa: phone works as web server
... head unit loads app and runs it

[ Structure of webserver on phone ]

sakazawa: phone: storage (media, web app, hmi), rpc, websocket server, http server, ssdp server
... Head unit (HMI, web apps (navi, maps, multimodal), browser (ssdp, devicce api -> CAN))

[ Relationship to functional requirements ]

sakazawa: this configuration satisfies some functional requirements
... contents + apps ubiquitous use
... storage access to content (media content is stored in phone, can be routed to head unit)
... native functions can be accessed from head unit
... web application in phone is personalized
... seamless networking
... storage area as cache/pre-fetch data

[ Demo ]

sakazawa: demo in Wireless Japan 2012
... article (in japanese)

[ Future activities ]

sakazawa: this is the wrong presentation
... marie will upload this slide
... we connect the smartphone and head unit
... head unit has navigation service

[ Future activities ]

sakazawa: standardization
... service discovery including privacy issues related to DAP WG (Web Intents)
... Create new value for users = Amazing Apps
... next generation context aware and personalized navigation
... multimodal apps + AI like a "Knight Rider"
... user could interact with the car
... that's our dream

[ Thank you. ]

[ Applause ]

Adam: questions?

takahashi: Takahashi, Mitsubishi Electric
... your structure seems similar to ACCESS's HaaP

marie: we uploaded your slides
... you can try again

[ Slides with picture ]

sakazawa: there's a smartphone and a headunit
... connected using WiFi-direct
... smartphone with restaurant destination

[ Personalization of appearance of headunit ]

[ Accessing smartphone content - from headunit ]

[ Buying theater ticket near destination ]

sakazawa: accessed at headunit using smartphone to website

[ Applause ]

Adam: 15 minute break
... we want to take a group photo after the break

Visteon: Web and Automotive

temmos: Jean-Marc Temmos
... Hello everybody
... yesterday I used a small script to compile and extract all the words from the position papers for this workshop
... the good news is that web and automotive are the biggest word
... w3c and api are smaller

[ Plan ]

temmos: a brief overview of Visteon
... HMI design challenges
... - tools and process
... - bring new features to the user
... Some hints for improvement
... - UI.Cockpit
... - Web technologies
... Expectations and recommendations

[ Visteon & HMI Innovation ]

temmos: Visteon is a Tier 1 automotive supplier
... we're (Visteon Software Technology, France R&D center) mainly focused on new generation HMIs
... for car manufacturers
... we're working on transversal frameworks
... we're involved with collaborative projects/academics
... concept car software design and development
... we work closely w/ German Center of Innovation
... we're an American company and have centers in the Detroit area

[ Visteon & HMI Innovation ]

temmos: Visteon Genivi Involvement
... we're on the Board of Directors
... etc.

[ Tools and Process ]

temmos: problems designing, from prototype to embedded target
... Many platforms/libraries
... lots of target specific code
... arm/atom/x86; Qt/SVG/Air; OpenGL/OpenGLES/OpenVG; Linux/Windows/QNX
... many tools, many design softwares: Flash, HTML5, EBGuide, Altia, Vaps XT
... many experts in various domains
... - artists, designers, developers - they don't want to use the same tools

[ Tools and Process ]

temmos: logics / graphics unclear frontier
... between behavior(intelligence) of application and the graphics side(way to display to user)
... relating to Tools
... HTML - CSS for skinning/displaying
... JavaScript
... Vaps/EBGuide can embed state chart
... there's no clear frontier
... famous gap between development cycles (consumer electronic v. automotive)
... consortiums: Autosar, Genivi - new opportunities

<scribe> ... new constraints

temmos: Technologies and standards change very fast
... Moblin -> MeeGo -> Tizen
... don't have to be early adopters, but have to follow
... E-Bee(?) with Renault
... we have to be as technology agnostic as possible
... we have to use as much web tech as possible - it's cross platform, easier
... the set of skills required is quite limited
... as compared to the low level tools

[ Improvements for the user ]

temmos: Customization
... Multimodal features
... context awareness
... open ended dynamic nature

[ Customization ]

temmos: Interface & personalization
... there's a Genivi group for it
... there's an API for detecting where a user is sitting, what they're doing
... car re-design for new usages (car sharing, renting, auto-driving...)
... User settings and preferences should be retrieved from the smartphone
... avoid "before google" mess for sync:
... customization is more than just skinning for css
... Facebook addict? want Facebook menu
... if i'm not, i want something else
... not possible with css

[ Customization ]

temmos: Left pedal, right pedal
... left for climate, right for infotainment

[ Multi-modal ]

temmos: many screens and displays
... haptics
... speakers & microphone
... - microphone array can detect position (is it a passenger), remove noise
... heads down display
... heads up display

[ Multi-modal ]

temmos: clusters & gauges
... off-board speech recognition in the cloud / onboard
... speech synthesis

[ Context awareness ]

temmos: context aware applications for infotainment and ADAS UCs
... location based
... map enhanced/map enabled
... onboard/offboard seamless switching
... - very good speech recognition in cloud
... - when i lose the system, i don't want things to be completely different
... cache some info to prevent degradation when the service provider changes
... driver attention, ... driver state inference

[ Open and Dynamic Architecture ]

temmos: we want the system to detect
... to support hot plug
... get content from car/phone

[ Open and Dynamic Architecture ]

temmos: seamless sync car->other (mobile, cloud)
... networking and web management
... integration w/ in car systems
... Machine to Machine (M2M) communication and Swarm Intelligence
... onboard/offboard traffic

[ O, Stream I/Open and Dynamic Architecture ]

temmos: in the past, software had to adapt to given hardware
... in consumer industry, consumer wants new game, buys new console, ... new model

temmos: resources are ...
... consider task
... with agents
... get to destination; ensure driving safely
... use modeling to derive hardware+software architecture

temmos: this is the task of the D3Cos project
... they use tools for this
... safety norms are converging now
... dynamic architecture to ensure safety for pilot and driver

[ Open and Dynamic Architecture ]

temmos: design patterns
... we learn a lot from the XXX partners
... provide us with multimodal human interfaces

[ E-Bee concept car ]

temmos: built using Linux+Genivi
... it's shown @Electronika in Germany (ending today)
... otherwise @CES

[ UI.Cockpit Framework ]

temmos: centralize application logic for cockpit HMI and distribute GUIs over various HTML5 client displays
... MVC based framework
... keep working w/ graphical environment that best fits on your target
... Altia might be a good choice
... but you might want to use TAT
... and you'd have to throw everything away
... client-server websocket
... xml api
... html5+js proxy

[ UI.Cockpit Framework ]

temmos: graphics, as many models as technologies

[ UI.Cockpit Framework ]

temmos: logic
... C
... or stateful C++
... or rhapsody
... whatever

[ UI.Cockpit Framework ]

temmos: how to fetch data
... CAN bus, MOST bus, ...

[ Expectations and recommendations ]

temmos: flexibility
... we face standardization process for ... and Genivi
... it was just standardizing what wasn't competitive
... there will be reluctance
... be flexible and have a limited scope
... low level access to device calls
... even if Android is getting better, for iPhone, we need more
... Graphics SVG - 2D/3G
... tools to compete w/ Adobe toolchain
... SVG is widely used, but not perfect
... clean logic / graphics separation

[ Video & more info ]

temmos: i'll show a video about E-Bee

[ Infotainment ]

temmos: btw, this was an HTML5 presentation using CSS
... no sound, i'll have to talk
... distributed system
... actually pretty centralized, we'd like more decentralized like UAV (unmanned air vehicles, i.e. drones)
... right side we have petal
... left petal is for ADAS, Climate
... usual frontier between Infotainment and ADAS
... giving screen can show both ADAS and Infotainment
... we should ensure every means of communication
... ensures safety
... there's no API for ADAS v. Infotainment

[ video crashed ]

[ switch to next video ]

temmos: ... augmented reality
... heads up / heads down
... context dependent hardware buttons
... this is a Nissan Leaf (for the E-Bee concept car)
... are some of you going to Las Vegas show (CES)?
... two camera in the car
... one in the back, one for QR code

[ last video: Driver ]

temmos: door modules for interact
... each passenger has their own module area with speakers
... improve configuration for each passenger
... 4 mics, 4 sound systems
... privacy on demand
... there are plenty of other videos on the web site
... thanks for your attention

[ Applause ]

dsr: time for one quick question

ph: All displays are driven by HTML5?

temmos: most of the display
... some are still in Flash

[ boo ]

temmos: i know it's bad
... but it's porting to HTML5
... we have Android tablets in html5

dsr: rich potential UI
... we'll pick it up later
... minor change to agenda
... i'd like to bring forward the Renault presentation

temmos: thank you

[ Applause ]

Topic: Drive the change - Renault

alidra: this isn't a presentation on R-Link
... I'm Ludovic Alidra
... from the advanced electronic project
... i'm involved in Genivi

[ Summary ]

alidra: car displays are special
... technical challenges
... web technologies and standards can help

[ car displays are special ]

alidra: with TV you have HMI for watch broadcast, play games, play dvd
... but you can still use it as a TV
... smartphone is similar
... place a call, use it to wake up, go to web, control tv
... but still a phone
... for a car!
... you could use HMI for radio, web, phone, HVAC/ADAS

alidra: but you MUST be able to drive your car

[ Safety first ]

alidra: not specific web usage in a car
... safety is already addressed by usual embedded HMI
... driver distraction management
... design rules, human factors, for app developers
... should not distract, stress or require driver action while driving
... validation of app before deployment on the app store
... until more dynamic run-time solutions can be proven

[ Challenge - Performance ]

alidra: user acceptance level
... response time requirements
... sw/hw optimizations
... high level for UX
... touch screen, speech
... animations, level of integration - feature coexistence
... native applications v. web apps
... html5 for native applications?

[ Challenges - Functional API definition (1/2) ]

alidra: both M2M and user-centric
... harmonized with Genivi
... API designed for:
... - performance
... - scalability and evolutivity
... low end products, high end products
... probably find a common api for basic elements
... each OEM wants to bring specific innovation
... need a way to extend API
... link to native application
... listen to web audio
... want to stop radio
... if you switch to weather app
... you want to still listen to web radio
... security/privacy
... always on-line
... need strategies to manage off-line mode

[ Challenges - Functional API definition (2/2) ]

alidra: examples of params for api
... remain connected (social network)
... audio video
... PIM (calendar, yellow pages)
... Help (BCall, HelpCall, ...)
... -- trip and car centric
... trip optimization (traffic, parking, ...)
... multimodality
... save money (eco-driving, PAYD, ...)
... - pay as you drive
... maintenance facilities (manual, after-sales offers, ...)
... - need status of your car

[ Challenges - Monetizing (1/2) ]

alidra: try to monetize your services
... one for application
... another for car data
... want to build our own app store
... collect/aggregate/distribute data

[ Challenges - Monetizing (2/2) -> M2M and User centric]

alidra: imagine we collect data
... traffic
... - location+speed
... provide to third party/special partner
... party adds element/creates services
... send back information to Car/connected app

[ Challenges - compatibility and interoperability with customer devices ]

alidra: not a new issue
... had problem for years, e.g. w/ Bluetooth
... gap between consumer electronics / automotive
... users expect same services in smart phone in car
... compatibility/continuity of services
... Car could be a node in Home network (DLNA)
... Mobile devices in car
... Car could be in cloud/C2Car

[ Web technologies and standards can help ]

alidra: improve quality/reduce costs
... better quality for end user
... more interop
... security - well known solution
... improved platform neutrality
... OS, HW independent
... improved code reuse
... application re-use and porting to new systems

[ End ]

[ Applause ]

ph: question... you said APIs
... have to be careful to standardize
... we want to compete on them
... to differentiate
... can you say a bit more about that?
... you have a whizbang feature in our car?
... want opportunity to expose that?

ph: why is that an issue?

alidra: you want examples of specific case?
... it isn't for me in this position
... if we standardize an api
... we'll have a small set
... very quickly, we'll need to add feature

ph: so, it's more you want the flexibility to extend
... so you can move forward w/o waiting

Josh_Soref: even at SDO slow pace, it seems we could standardize before you ship

alidra: if we use a standard, we'll win at the end
... we'll reduce duration of development
... if we have our own api, we'll start to launch application
... we'll win later when it's standardized

dsr: thanks

[ Applause ]

TOPIC The Architecture of Future Automotive Applications based on Web Technologies

speiser: Sebastian Speiser
... we're not specific to cars
... we're looking at cyber physical systems
... i learned a lot of new car terms in the last days

[ Motivation ]

speiser: not just getting B from A
... driver has other goals

[ Motivation ]

speiser: doesn't have any heads up
... because he has speech and big team
... Formula 1 spends 1k USD / km
... Germany gets 30c / km

speiser: speech recognition, virtual knowledge based integrating data from the web
... logical inferences, statistical learning

[ High-level Architecture ]

speiser: driver -> speech i/o
... -> memory -> Interoperation, Data I/O
... Dataa I/O -> LOD
... Interoperation -> Logic Reasoner, Adaptive learner

[ Overview of Architecture ]

speiser: for speech I/O
... don't want ...
... use patterns to convert text to logic
... use reasoning engine to control conversation
... using Cyc's project to link logic to language
... Stream I/O
... also web streams
... for temp/traffic

[ Data I/O: Linking Open Data Cloud ]

speiser: in 2006, Linked Data movement
... semantic web in W3C

speiser: yearly growth rates of 200%
... small statements

scribe: very large set of connected data sets

[ Combining and using I/O channels ]

[ Memory: Virtual Knowledge Base ]

speiser: based on Linked Data and REST principles
... basic abstraction: a resource
... POI as 3D XML, RDF or JPEG
... resources have references (links) to other related resources, e.g., leading to the next step in a series of operations
... support standard operations: CRUD
... "Colosseum is in Rome"
... Create-Read-Update-Delete
... I'm not a REST-afarian
... beyond scalability
... you have standard operation
... interop is easier
... common operations
... fixed grammar
... just standardize vocabulary
... very different from WS*
... which was ported to HTTP and called REST
... You do a HTTP lookup on URI
... do it with an RDF program and get a reply

[ Interlinking in Linked Data ]

speiser: car:currentDriver owl:sameAs facebook:JohnDee

car: car :location car:point ... wikipedia

[ Streams, Services and Compositions ]

speiser: streams either just leave http open and continue to listen
... or pull stream source regularly
... Services, to integrate w/ linked data
... how to build applications, compositions, workflows
... based on REST resources?
... From university perspective, this is established technology

[ Linked Services ]

speiser: services honor REST
... we wanted to use RDF/Notation3
... relation between input and output
... linked between services, descriptions, and data input/output

[ Production Rules for Linked Services ]

speiser: my colleagues are presenting a paper about data driven programming for dynamic programming
... evaluate rules, if it matches, invoke action(service)
... Locate() -> car location -> Lookup() -> list of gas stations -> navigate() -> programmed GPS system -> done
... car location -> cancel() -> new location -> done

[ Conclusions ]

speiser: use web tech to provide assistant system to drivers at reasonable cost
... make it cheaper by not needing so much adaption for every platform
... web tech in front end + back end, we can add new sources on demand
... PAYG integration
... no need for up front specification of services
... tap long tail of sensors and data sources

[ Thank you for your attention ]

speiser: have a loook at www.kskri.kit.edu + www.aifb.kit.edu
... we have a smart energy grid where we're using some of our proposed technologies

[ Applause ]

dsr: Questions?
... idea to combine this w/ semantic web
... what's needed to build an ecosystem?

speiser: in terms of what?

dsr: what's required to bring to market?

speiser: dev cycle in car are rather long
... intermediate step is probably tablet/smartphone applications
... and installing this in a car
... we have cars from Daimler, Opal
... maybe a car manufacturer would be interested and take this on

dsr: thank you

[ Applause ]

Proposing a Meta-Language for Specifying Presentation Complexity in order to Support System Situation Awareness

feld: Michael Felt, German research center for AI, UI group, Automotive Group
... we're building HMIs for driver systems
... do a lot of user testing
... we do this in conjunction with OEMs

feld: this work is related to situation awareness

[ Situation Awareness ]

feld: knowing what's around you

feld: went from psychological model
... define system situation awareness
... started with Endsley Model

[ Endsley Model (original) ]

feld: three levels: perception; comprehension; projection

feld: i see a car
... i'm going to crash
... i need to press breaks
... there's a feedback loop
... there are task factors

[ Endsley Model (adjusted) ]

feld: assessment of user+context; update information sources ; impact estimation
... intelligent mediation
... HMI: presentation of information

[ Two-fold Research Question ]

feld: driver aspects / presentation aspects
... define system situation awareness
... driver related concepts / presentation task annotation

[ Estimating presentation complexity ]

feld: we have 3 options
... complexity from designer
... ideal case -> nothing to do
... (unlikely)
... unstructured representation of HMI
... Image
... difficult to do something
... could use heuristic, approach, do image processing
... -> low confidence
... structured representation (HTML5)

feld: "ACE" (annotated complexity estimation)
... how do we annotate complexity automatically
... ACE is based on visual tree of interface

[ Example Screen Layout (sim TD) ]

feld: based on a German collaboration
... German Options screen

[ GUI Model and Visual Tree ]

feld: from representation, we can derive visual tree

[ Complexity Computation ]

feld: compute it based on a table
... label, icon, panel
... we assume that if we do the right studies
... we can get a sufficiently close approximation of a good complexity

[ Presentation Meta Language ]

feld: developer can specify multiple alternatives w/ different complexity
... system can choose based on situation
... we didn't want to invent a presentation language
... so we suggested a wrapper/meta language

[ it isn't well formed XML either ]

[ Implementation info a Dialogue Platform ]

feld: SiAM dp (Situation-Aware Muiltimodal Dialogue Platform)
... based on SCXML state charts
... it can decide based on different ways to present something
... can do this online or offline
... we can check how high/low the load is for a dialog

[ Estimating (Input) Interaction Workload ]

feld: we have to quantify user interaction cost as well
... break up complex tasks into atomic tasks
... rearrangement, text entry, number entry, pan/zoom, list selection, scrolling
... -> widgets
... i'll probably build a similar table
... to presentation complexity
... to know what the cost of a certain tax in a modality costs

[ Two-fold Research Question ]

[ Driver-related Cognitive Load Aspects ]

feld: how to model cognitive load?
... how to quantify cognitive load?
... there's no such easy measure
... no "cpu cycles for human"
... hard to use time values

[ Cognitive Load Model ]

feld: top down view of literature concepts
... driver interacting w/ system
... demands "dry"
... specified in table
... cognitive demand
... environment/user interactions
... cause cognitive load
... assume multiple resources available to human
... possibly not completely separate
... all cognitive demands are transformed for the different processing resources
... we have stage/dimension/code

feld: maximum amounts of capacity differs for different people
... looking at levels of fill is the current cognitive load
... metrics can detect stress/distraction and affect load capacity

[ Cognitive User Model ]

feld: user
... processingresource (1...n)
... dimension
... cognitivecapacity
... cognitivecost (1..n)

feld: dimensions
... interaction

[ Related Efforts ]

feld: two more topics

[ Automotive Ontology ]

feld: where do we store this
... yesterday we heard about sensors/knowledge in cars
... have to store it somewhere and make it available
... we created an ontology
... published last year on automotive ui

[ High-Level Structure ]

feld: from a user's perspective

[ User Model ]

feld: it could be worth adding user dimensions too
... our model is based around UML
... with mental state, time pressure, emotional state

[ Meta Information ]

feld: time, confidence, location, privacy
... sensors are more or less accurate
... privacy for Car-to-Car communication, what information can we share?
... location should be annotated for most values

[ Driving Simulation ]

feld: driving deviation
... is how do we measure cognitive load
... in European funded "Drive Home Safe"
... driving simulation connected directly to HMI
... sign on road
... trigger signs
... there's a physics engine

[ Driving Performance Measures ]

feld: we can derive performance metrics
... we can add further driving tasks
... different UIs require different tasks to test
... not just lane change
... we're looking into standardizing work tasks

[ Traffic Light Control ]

[ External Visualization ]

feld: we worked w/ intervisual-computing-center
... so you can swap engine

[ More Information on OpenDS ]

feld: thank you for your attention

[ Applause ]

dsr: thank you, that was interesting
... do you think it'll be easy to provide tools for developers to test?

feld: that's part of the Scion project
... it's an important point

dsr: the car is trying to understand things that are coming

feld: we have a knowledge base
... we're working on adaptivity strategies
... UI impls or interaction strategies

dsr: have you tested with driver state?

feld: we did Multimodal comparisons
... with biosensors
... and saw effects there

kaz: curious about protocol between modalities
... and SCXML
... do you use MMI lifecycle events and EMMA?

kaz: between simulator and speech and touchpanel?

feld: it's a proprietary interface built into SimTD

dsr: thanks so much

[ Applause ]

DSR: time for lunch ... until 2pm

marie: note that there are participants lists at the registration desk, pick up a copy if you like
... note also that slides are on the web (linked from the agenda)

Adam: Our next speaker is a colleague of mine from Intel. He has experience from Nissan,
... I want to introduce Nobe-san from Intel

Nobe: Before March of this year I worked for Nissan and did many implementations.
... I have good experience on how ICT companies can work with auto companies.
... Topic is Connected Car as a Cloud Computing Client.
... The implementation will only be for 25% or 30% of the autos.

[What is a Connected Car]

Nobe: There is a control area that has been implemented for 60 years.
... But there is instrumentation in the IVI that has been around since 2008.
... These have been monlithic designs.
... Smartphones have made it possible for these devices to be generic.
... Of course there should be some connection between a IVI and control areas.
... Data in Control Area should be read only.
... The connection could be a smartphone or tablet inside the car. Some kind of connection device.
... Make it easier to provide information to the driver.

[slide has been built out]

Nobe: Some is sensor data and some is Big Data supported by Cloud computing.

[Issues to embed ICT devices in the Car]

Nobe: We have to reconcile the difference in development life cycles.
... Development cycle is 3 years followed by 3 years of sales.
... The average car stays in the market for 7 years after purchase.
... There is a total of 17 years that a vehicle design is current in the market.
... This is the conceptual difference between a vehicle and ICT.
... If we're going to embed a communications model that will change during the life of the car.
... Current models may be 2G based but next generation will be different.
... We need to think of what the requirements will be 7-8 years ahead.
... The example is on the MIPS performance available compared to the time.
... Then we have to look at computer usage for the next 10 years in the market.

[Issues to embed ICT devices in the car ]
[Application update will solve the problem]

Nobe: We have to update and download applications on a regular basis.
... Hardware is completed and fixed, but applications need to be updated.
... This concept is important to remember as we develop applications.

[Market Penetration of NAVI: Recnet Change]

Nobe: After introduction of iPhone the market acceptance has been greatly accelerated.
... Will be 1B units by 2015.

[Contents and Services reside in the Cloud]

Nobe: Most of the functionality supported by PCs will be supported by Smartphones.
... And the car will be connected to the Cloud via the smartphone.

[Forseeing the Future Trend]

Nobe: Three factors: Displays, Cloud and Smartphone.

[How Smartphone connects Car with Cloud]

Nobe: Smartphone migrations will change every 2 years, but the car system will be in place for 10 years or so.
... We need collaboration between the OEM and Smartphone providers.

[Role of HTML5 for IVI Client]

[discussion on how HTML5 and associated standards can work with car]

Nobe: I think it's important that the car use HTML5 even when it doesn't have access to a server.

[Use of Internet Contents using TA (2006)]

Nobe: This is a little future look, but it's also what I was doing at Nissan.
... It links Semantic Web with the car.
... With the information, the vehicle may send out a request about how it can be serviced at a gas station.
... The display can show the information that's important for the vehicle.
... The navigation system should only display what's important for this vehicle.
... HTML5 should be implemented with the semantic types.
... Need the Web to support this with Web Services.
... It's been implemented and is being used in Japanese market.

[TA: Telematics Agent slide]

[discussion on workflow / dataflow]

Nobe: Displays things like name of station and directions to get there.
... This was done fairly easily.
... It's a Web based implementation.

[Example of Big Data Applications]

Nobe: Location information is time-stamped. It can provide information on the traffic conditions.
... This kind of information can be used for traffic jam avoidance.

[Crisis Response on Google Maps slide]

[Possibilities to apply Big Data Analysis slide]

[skipped 2 - 3 slides]

[New Mobile Service Opportunities with Big Data slide]

Nobe: This data could be used by other companies as made available by third party vendors.
... Could offer new products and services.

[Summary slide]

Nobe: Cars will be connected.
... HTML5 is the "glue"
... for OEMs the ICT implementations require horizontal collaborations

[applause]

Roger: Your previous employer and Toyota are both in danger of having fragmented solutions. Are they looking to solve that in the future?

Nobe: I think there are services that will be the same, but the implementations have been different.
... They need to get to a common base which would require hardware and software to be abandoned. I think the market will drive that.

dsr: You mentioned at the begining that the design of cars takes time, do you see that changing in the future?

Nobe: Things are getting more modularity, and that implementation may make it quicker.
... The cycle is 3 years so you would have to change the sales cycle to match it.

[Applause]

Acheiving a Safe and Seamless UX through accessible web apps

iglesias: Alejandro Pinerio Iglesias, Igalia
... this year i started working on WebKit
... with ARIA and that stuff

[ Topics ]

iglesias: Acessibility-based solutions to improve safety
... implementation via accessibility APIs
... applicable standards and existing support

[ Acessibility-based solutions to improve safety ]

[ Safety First? ]

iglesias: we have several cases of problems related to distraction and accidents
... people texting and walking have an accident
... in 2010, 3092 people killed in accidents related to distracted drivers

[ Hands-Free via Speech Input ]

[ Eyes-Minimal via Simplified UI ]

iglesias: a minimal simplified UI
... is useful for people w/ difficulty with screen contents
... but also people who should not be distracted

[ Eyes-Free via Gestures ]

iglesias: users who cannot activate on-screen elements (disability)
... users who should not activate on-screen elements (driving)
... shouldn't use touchscreen while driving, shouldn't look at screen while driving
... shouldn't be distracted by complex UI

[ Studies ]

iglesias: driving is primarily a visual interaction w/ road

[ Implementation via Accessibility APIs ]

[ Accessibility APIs ]

iglesias: accessibility APIs allow you to interact w/ apps programmatically on behalf of user
... use Voice/Sound

[ Example: Remember the Milk ]

iglesias: driver goes through an area
... Remember The Milk has an alert for the area
... device checks speed decides it's safe to speak the alert
... driver says "No" to indicate response
... device clicks "No"

[ Accessibility API Events ]

iglesias: Focus changes
... selection changes
... text changes

[ Accessibility API Actions ]

iglesias: press, release, click, ...

[ Accessibility API Interfaces ]

iglesias: application, component, desktop, document, ...

[ Accessibility API UCs ]

iglesias: access to users with disabilities
... automated testing
... distraction-free access?
... most testing is probably Unit Testing
... but you eventually need to write Functional testing
... Accessibility is a good way to do Functional testing

[ Requirements ]

iglesias: developers: create accessible web apps
... layout engine implements accessibility support

[ For Web Application Developers ]

iglesias: W3C Mobile Web Initiative
... Best Practices

[ For Web Application Developers ]

iglesias: W3C Web Accessibility Initiative (WAI)
... Accessibility Overview
... Guidelines
... ARIA

[ For Web Application Developers ]

[ Layout Engines Supporting ARIA ]

iglesias: Trident
... WebKit
... Gecko
... Presto
... Trident = IE
... WebKit = most others
... Gecko = Firefox
... Presto = Opera

[ WebKit: Free as in Freedom ]

[ WebKit: Embedded-Friendly ]

iglesias: Desktop (Safari, Chromium, Epiphany, )
... iPad, iPod, iPhone

[ WebKit: Developer Community ]

iglesias: it's hard to quantify people
... there's chromium, apple, webkit, google, igalia, gmail, nokia, openbossa, samsunbg, rim, others

[ WebKit: Accessibility ]

iglesias: mature, platform-independent
... adaptations for each platform/port
... Macintosh, Chromium, GTK <- mature
... EFL, Qt, Win <- less mature

[ The Indie UI WG ]

iglesias: Accessibility/Automotive have the same targets
... they're working separately

[ Indie UI WG Members]

iglesias: access, apple, google, nokia, opera, ...
... but no automotive members

[ Summary ]

iglesias: in-vehicle applications can be used more safely via alternative input/output
... quite similar to assistive technologies
... accessibility APIs are standards for accessible web applications, established, used
... automotive industry should evaluate this solution

iglesias: - the need relating to alternative input/output
... is where the accessibility projects have been talking for years

[ Questions? ]

[ Applause ]

dsr: cars have a rich set of UI possibilities
... i don't know how wide the range of the INDIE scope is
... headsup-displays, spatial audio? tactic?
... multidevice things

iglesias: they need text input
... they need text output
... screens with icons
... this isn't acceptible in HMI
... for an accessible interface

kaz: there was discussion at Indie UI during TPAC for UCs
... i was wondering if you're interested in giving feedback about feedback
... from ecosystems
... to the Indie group

iglesias: iPhone uses "Voice-over"
... they rely on gestures

Adam: thanks iglesias

[ Applause ]

The Connected Car-Changing Business Models for Automotive

delaney: Chris Delaney, CSC

delaney: a colleague put together a position paper
... i changed our presentation
... to talk about sales and deriving value

[ Agenda ]

delaney: Driving in the Cloud
... impact of HTML5 on IVI
... new technologies, new business models

delaney: streamlining sales process
... delivering a customer for life
... vision of the future

[ Driving in the Cloud ]

delaney: left, what happens when you're building a car
... right, what happens after you've sold the car
... often there's no interaction between seller and buyer for 5 years

[ Impact of HTML5 on IVI ]

[ new technologies, new business models ]

delaney: streamlining sales process; engage w/ customers
... buy something from Amazon, very streamlined, simple process
... buying houses hasn't changed in years
... same for cars
... enable customers to customize product on the fly
... enable fast delivery
... let customers know where the car is
... integrating Facebook w/ Salesforce, SAP
... deliver a customer for life
... deal w/ them on an ongoing basis
... McKinsey: most companies have improved productivity and operations, but not sales

[ Pending Business Model Transformation ]

delaney: Spotify fans?
... much more advanced than CDs
... subscription model v. purchasing model

delaney: discussed Zip cars in SF, CA, US over lunch a few days ago
... renting cars instead of buying

[ streamlining sales process ]

[ current sales process ]

delaney: you can go on internet, do research
... there may be a car configurator, In Flash
... maybe you can book a test drive
... for further info, contact dealer
... backwards/forwards negotiation
... not like Amazon

[ Engage with customers: Seamless customer for life Communication ]

delaney: encourage customers to return to OEM for everything

[ Demo Scenarios ]

delaney: 1 - Purchase Process
... 2 - In Life Management
... the idea is to generate discussion

[ Potential Customer Interested in Honda - Likes Honda Facebook Page ]

delaney: customer likes facebook page
... could have a lead in salesforce
... have 1-1 dialog

[ Customer Downloads Honda App ]

delaney: encourage users to download it
... through iTunes store/etc.

[ App Integrated with Facebook ]

delaney: viewed at home through your iPad

[ App start-up ]

delaney: find showrooms, book testdrive, etc.

[ Customer Books Test Drive at Local Dealer Using App ]

delaney: pick model, pick date and time, confirms details, confirmation

[ Customer Checks Availability of Models of Interest ]

delaney: integrate model info w/ SAP data
... maybe OEM is willing to do a deal with a unit in a different location

[ Customer Configuring New Car ]

delaney: this is just reusing content from web

[ Stock data ]

delaney: pulling data from SAP

[ Payment Options - Buy Now (Cash) or Finance ]

delaney: go through buying process through app
... credit checks, finance - online
... give feedback on process

[ A New Honda Advocate ]

delaney: then you can post to Facebook
... great free advertising
... @home, i'll look at reviews on Amazon
... how do we leverage the power of social?

[ Scenario 2 - In Life Management - My Honda ]

delaney: they bought car
... find out about servicing, hooked up to telematics
... you need to change oil
... special offers
... make it useful
... they don't want spam

delaney: if it's targeted, thought about, a value to consumer, no problem

[ Promotions and Special Offers ]

delaney: that top part is from Salesforce

[ Customer Details Shown in SFDC ]

[ A vision of the future ]

delaney: HTML5 is what our app was based on
... communicate 1-1 w/ customers
... OEMs say they don't know much once it's sold
... having them come back for servicing/other products...
... Zip-car, where people rent the car
... there's an opportunity to communicate more often
... this applies to a wide range of manufacturers
... in UK, the decline of High Street stores
... producer manufacturer to customer relationship gets stronger

[ Questions ]

[ Applause ]

takahashi: from your PoV
... what kind of standardization is needed for Automotive?

delaney: this can be done through a wide variety
... the honda app was built using HTML5
... it could be built w/ anything
... Facebook - Salesforce integration is very good

Adam: thanks delaney

[ Applause ]

Decisions on Future Work

ph:Expect preliminary results in this session only.
...We expect to continue this conversation to finalize
....Outline of what we heard and why W3C is the place to work
....Also when Standardization works well.

[Motivation for Web Technology slide]

ph:Seens strategic for industry, there are demands and expectations.
...There is demand there. We need to build an ecosystem of developers
....There isn't a huge market per OEM, but need to build industry demand
....Are there other things missing?.

DanA:In addition to work at W3C there is lots of other work being done. Can W3C serve a role to harmonize these efforts globally.

[Motivation for W3C slide]

ph:Historically there is a problem in working on Web standards in isolated community (Mobile, TV).
...Developers are watching what W3C does.
...And W3C standards are Royalty-Free that makes them easy to implement
....Maybe I missed some, if you have other motivations, let me know.

[crickets]

Boris:We heard about connectivity to the cloud. We hear W3C sits there. That might be another motivation.

[Standardisation: How does it work Well? slide]

[Prioritize Work Items]

ph:I try to collect work items from the audience
....The idea is that people say what are the things W3C should work on
...I'm personally interested in the vehicle APIs
...But it would be good to hear from attendees what they want and some reasons why.
...This is preliminary but it gives us a feel for what was represented.

andy:By M2M you mean telematics?

ph:I heard someone say it, I just wrote it down.

andy:The first five that are on your list are of interest to me.

ph:OK - we'll move those to the list.

andy:I would define user-centric as a use case dealing with the driver or passenger. M2M is vehicle data for things like insurance / fleet maintenance, etc.
...Could be anything you would use to design an API.

boris:So entertainment API would be part of that as well.

hirabayashi:I think connection in vehicle. We have to connect head unit and smart phone.

andy:If you consider that to be a user-centric device then I would say yes
....I don't think those names are what we would standardize on, we'll end up with sublists
....Something like device-focused APIs are underneath that category as well.

ph:So device centric is interfacing smart phone.

mathias:So I think we need to think cross platform.....How to integrate the HMI..Having something on top of HTML5.

dsr:Was this something we heard earlier about application portability>

pavel:I think we should take out API. I think there are things deeper than that....We might not want to partition APIs this way....There might be overlapping APIs in the categories....We probably don't know details today....I would not say the APIs will need to be partioned the way of Use Cases.

andy:To me that makes sense. In terms of strictly bidding them, from an application point of view, they might overlap. We have a virtual mechanic app like that....I'm willing to wait until a later time to categorize them.

magnus:Seeing the list and the concepts. I think people maynot be on the same page.
...Perhaps we develop an abstract reference model where you can put concepts and APIs in the same picture
....It might facilitate discussion.

kazuhiro:Some of the speakers talked about ecosystems.
...Right now we don't have any energy for ecosystems.
...W3C might help build that and make it happen.

Roberto:The way we follow it in Genivi is we list the Use Cases etc. across competitive things. Many companies are uncomfortable with such discussions
....Trying to see what's available on the car and putting together networks that are available, then putting APIs that are available for building apps
....This is a different approach but it has delivered results for us.

ph:I share your opinion and approach, but others feel it's important.

woochul:We need to figure out what's needed as automotive API. What's missing, what's reusable.

ph:That sounds like gap analysis. Think Vodafone took a shot at that.

magnus:I think looking at the board, we have priorities to discuss. If you have different applications running how do you prioritize them...That should be included in here.

dsr:I think that's what context is. Based on teh situation, what's the priority.

magnus:I also think we should consider how to create a developer society so tooling might be an important part of this.

ph:What do you mean by tooling?

magnus:Provide guidance on how people can build things without an actual car....Maybe best practices.

??:I was going to mention best practices as well.

mayer:I think we should mention something about long life cycles and synchronization as well.

??:When it comes to lifecycles and such I think OMA is doing work there so we should work with them.

woochul:I'm still not sure of what level of API we should consider....Low-level API could be hard as each car manufacturer has it's own implementation.
...An extraction layer would be easier.
...I'd like to know how to work on such issues.

ph:We can put up both topics and see where they prioritize.

Josh_Soref:I think we should look at both (where individual vendors work with a few 2nd parties to play with low level apis to see what kind of apps they can build to generate future high level UCs) and then figure out who will work on which ones and have people focus on the Use Cases they could implement around them.

phShould we get a show of hands on the topics? I would ask that you only vote 3 times....Will do next round probably through an online questionaire.


[Voting by topic]

Ideas and counts in descending order

ph:The votes are in, thanks very much, we'll publish this and see what people what to focus on.

[How to get Started? slide]

ph:W3C submission from a set of members.
...We do Working Groups which is where we do the body of our work.
...Where requirements arse done and work is created.
...Liaisons with other groups.
..And it's member only work.

ph:An intermediate step we've introduced is Busienss Groups which doesn't normally develop standards
...It is designed to work with other groups to figure out what to do next..
...Just wanted to give you this overview and have you think about it as the way forward.

ph:Wrapping up I'd like to thank Intel, and to our Sponsors QNX and Webinos project.

<Alan> [big applause]

marie:Many thanks to all the scribes and most specifically to Josh Soref!

<Alan> [adjourned!]

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.137 (CVS log)
$Date: 2012-11-30 12:18:15 $