W3C Augmented Reality Workshop (Day 1)

15 Jun 2010

See also: IRC log


See Attendee list
Dan Appelquist (DKA), Phil Archer, Christine Perey, Rittwik Jana
phila, Phil, Rittwik, Dan, francois


See the Agenda.

  1. Setting the Scene
  2. Framing the Question
  3. What Role for Standards
  4. Demos

<DKA1> Hi

Setting the Scene

<phila> CP: opens workshop

<phila> Thanks Manel Medina for hosting the workshop

<phila> Thanks W3C

<phila> DKA: Introduces himself

<phila> Rittwik: Introduces himself

<DKA> Scribe: phila

<DKA> ScribeNick: phila

Phil: Introduces himself

CP: Introduces herself & workshop [slides]

CP: Talks about the emergence of AR
... 'AR Continues what the Web began'
... Focus will be on tech here but also need to think about business, sociocultural issues and legal/ethical stuff
... lists current limitations of AR
... asks to hold off on discussion of applications vs. browsers as that will be covered in Mobile 2.0
... not starting from zero. Refers back to AR Summit at MWC in February

CP: Promotes ISMAR in Seoul later this year
... proposes an AR Ecosystem
... Content box is big
... that's where the money is?
... user has a device with lots of sensors
... may be adding to the social meda data
... companies want to connect with those users
... then there are enabling technologies and that's the focus of the workshop
... sensors -> triggers, some of which use existing standards
... content provider needs to specify what the trigger will be for what
... hands over to Damon Hernandez (DH) to talk about existing standards

DH: Introduces himself from Web3D
... and other companies
... Wears lots of hats
... based in San Francisco and was at ARE
... event was focussed on commercial side of things. mateo, Total Immersion etc. present
... our panel on Next Gen AR
... we talked about the way the different tech can come together
... very GeoSpatial based
... talked about the georgia Tech KAMRA app

'Going forward' is a key slide

scribe: OGC has about 30 different standards
... including KML
... CityGML is a biggie too
... web3D consortium. X3D is the new standards that supersedes VRML
... used for 3D Web apps
... and 3D TV

The world is our content the Web is our platform

scribe: how do we integrate with real world companies with existing content
... 2D data incl. geo located tweets, Flickr pics etc.
... 3D is very important. Necessity to create once and display many

<DKA> ScribeNick: DKA

Phil Archer intro to W3C [slides]

Phila: Many organizations already involved in standards space already listed by Damon.
... my job at this workshop is to understand where W3C should fit in.
... we have a process we have to follow in W3C to create recommendations.
... everything we do is done because the members wish it.
... e.g. HTML5 came about because W3C members wanted it to happen.
... Paying members are influential.
... however - anyone in the public can participate and be a part of W3C standards process.
... we have a process where we must reply to all public comments.
... if you're in a small organization, you can still take part. We have an invited expert process for people working for themselves or for small organizations.
... [presents w3c process]
... [how a document goes to recommendation]
... [presents the incubator group]
... it is easier to become an invited expert on an XG. There may be scope to form an XG [on AR] based on this workshop.
... This workshop is partially funded by the OpenMediaWeb project - thanks to the European Commission for that.

<scribe> Scribe: Phil

<scribe> ScribeNick: phila

CP: returns to her AR Ecosystem

Tour de table

Alex from IBM

scribe: works in consulting organisation
... developing apps for things like Wimbledon
... working with folk like DKA

Timo from Fraunhofer Institute

scribe: does graphic and data processzing
... interested in visual computing and interaction

Nzubke - interest from academic side. I'm doing PhD. Looking at healthcare AR

<phila_> realised I'd been typing into nothing... Grrr

<phila_> Missed some of the intros

<phila_> Telecom Italia Marco and Claudio introduce themselves

<phila_> Manel Medina from UPC

<phila_> Klas Harmondsson from Sonyericsson

<phila_> we think AR is an interesting way to provide new experiences for our users

<phila_> says nice words about W3C

<phila_> Dirk Groten from Layar

<phila_> Layar is closed, few standards

<phila_> we have 2 million users so maybe it's time to think about standards

<phila_> Wolfgang Damm from Mobilizy/Wikitude

<phila_> Tomasz from Telefonica R&D

<phila_> Vincent Reynolds from DERI

<phila_> A Sem Web centre in Galway

<phila_> Take work done in DERI and apply it to mobile world

<phila_> AR is obviously an area

<phila_> Sunghan from Korea (ETRI)

<phila_> Fatima from ARPA Solutions

<phila_> commercial co making products for various manufacturers etc.

<phila_> Martin from Uni Oviedo

<phila_> Working on human computing etc

<phila_> Margherita from EC. Does research and interested in mobile search

<phila_> Jonghong from ETRI

<phila_> research interest is next generation Web

<phila_> AR not my main area but Web and AR look like being combined in the future

<phila_> Martin Alvarez from W3C Spain office

<phila_> David from The Young Generation, an ad agency in Amsterdam that uses AR

<phila_> Simon also from The Young generation

<phila_> Pascal from Total Immersion

<phila_> Josep from Barcelona

<phila_> ... involved in Mobile Monday here

<phila_> Johan Dupuis

<phila_> ... describes his project working with 3D objects

<phila_> Some guy from W3C called Francois introduces himself

<phila_> Damon H

<phila_> Jordi Janner from local university, working on soundscape generation

<phila_> Jacques Lemordant from INRIA

<phila_> Jose gato from Uni Juan Carlos from Madrid. Working on open source software (Libre Soft)

<phila_> Raul from same uni

<phila_> BDigital

<francois> ScribeNick: francois

Phila: [mentioning that minutes and report will be public]
... [counting people coming to diner: 26 participants raise hands]

Framing the Question

<phila_> DKA: says some opening remarks. Reviews the range of people in the room

<phila_> big geographical spread

DKA: Really interesting to hear all of the interest that people gave in this first round. The ecosystem is here, obviously, service providers, operators, manufacturing actors, advertising agencies, consultants, academia, business companies, from different regions in the world.
... If nothing else comes out of this workshop, it should at least allow to create links between these different actors.

<phila_> Wants to talk about importance of Open

DKA: Also feel free to come to Mobile 2.0 conf on Thursday for more networking possibilities.
... I remember the dark times of pre-web technologies such as gopher, usenet, pre-web services (AOL, Compuserve, ...)
... I talked with my 6 year-old daughter about the Web, saying that Tim Berners-Lee invented the Web.
... She said "What Web?"
... The Web is like air for her, it has always been there.
... Augmented Reality is not mainstream yet, but I feel we're at a tipping point.
... There will come a time when users will just use AR and not talk about AR.
... Part of the necessary conditions that lead to that state are openness, standards.
... The Web wouldn't be where it is without that.
... Leaving the floor to Jonghong Jeon from ETRI.

Jonghong's slides

<phila_> I am not making a concrete proposal

<phila_> Offers a definition of AR

<phila_> Shows good picture of himself as Oliver Hardy

<phila_> 3 kinds of technology

<phila_> 1. Contextual awareness

<phila_> info architecture

<phila_> visualisation tech

<phila_> history of AR

<phila_> Starts in 1970

<phila_> Lots of developments in later years

<phila_> smartphone tech made progress in last 10 years

<phila_> 2 types of Ar approach

<phila_> 1. Visualisation approach

<phila_> 2. Informative approach

<phila_> Shows slide with differences between the two.

<phila_> Informative approach gives 4 types of applications. (Shows slide)

<phila_> Most important software is the AR Browser

<phila_> Slide shows more detail

<phila_> AR on the Web?

<phila_> Take a live video as 'reality'. 2nd is augmented info. 3rd part is the Web itself

<phila_> Poses a series of questions on a slide

<phila_> Shows general framework for AR

<phila_> Why do we have to another browser?

<phila_> Offers Pros and Cons of extending regular browser to support AR

<phila_> (i.e. doing away with concept of separate AR browser)

<phila_> Suggests 12 issues to be solved for interoperability

<phila_> formats

<phila_> transport and interaction

<phila_> 3. representing 3D interactive and mixed reality contnet

<phila_> 4. Event description engine

<phila_> 5. Local caching method

<phila_> 6. Additional functionality of UA

<phila_> 7. device capabilities

<phila_> 8. AR data mash up method

<phila_> 9. AR data format

<phila_> POI - how can we describe POI info? social info?

<phila_> 10. AR data service API

<phila_> 11. Open markers DB issue. How can we use common marker databases?

<phila_> 12. Security and privacy

<phila_> Intersting Web Architecture technology stack diagram

<phila_> HTML 5 at the top, URIs and HTTP at the bottom

<phila_> What kind of standards required for AR on the Web?

<phila_> Blue - means current standards

<phila_> Purple - current work in process

<phila_> Interaction/Gesture API is a future work item?

<phila_> Then more future work at top of stack

<phila_> Wrap some of these standard in HTML 5?

<phila_> Concluding remarks on slide

<phila_> Need a new working group

<phila_> to develop AR on the Web

<phila_> Need a live video streaming standard in HTML 5

<phila_> Items in red are new items we need to look at

<phila_> And the two below then

<phila_> More work items under way in 2nd conclusion slide

<phila_> ETRI makes commitment to support W3C activity in area of AR

<scribe> ScribeNick: francois

DKA: Thanks, Jonghong, each time I hear you talk, I get an insight of the particularities of the Korean market.

JJ: many companies in Korea think that AR is the user interface of the future.

phila: I'll be presenting the paper on Augmentation Concerns in the absence of his author. I didn't write this paper, obviously.
... Hopefully, I won't deviate from the author's intent. [slides]

<DKA> [JJ noted that Samsung and LG electronics were building AR services and embedding these (applications?) into their smartphones for the Korean market]

phila: Everything that's in AR already exist in some way, so concerns that generally arrise on other topics also arise in AR.
... Concern 1 is privacy. If a vending machine knows whether you prefer Pepsi or Cocal-Cola, should it serve you the drink you prefer, should it display an ad targeted at you?
... Concern 2 is context. How can you differentiate the different platforms you share info with (LinkedIn, Facebook, Twitter, Flickr, …)
... Concern 3 is recognizing individuals. You may well share info without thinking that this info will be used in another context. Example of drinking a beer in a bar. Should that be used everywhere in other situations?
... It's all about identity control. For instance, I'm very protective of my children, never tweet their names or publish a picture of them.
... But I feel the social pressure to publish pictures of them on the Web.
... Paper trail is another thing: is it right that a restaurant in Barcelona can stat the number of British people that enter it? In some ways, it may help them adjust their menu, in other ways, it's a breach of anonimity.
... Final concern: enhance, don't detract!

DKA: next on stage is Dirk Groten from Layar.[slides]

(Note, Romain Bellasort unable to attend at this point due to severe travel delays. He would have presented these slides.

<Rittwik> Scribe: Rittwik

Dirk Groten, Layar to speak on "Components of AR as a mass medium: A view on standardization"

Layar is pioneering on AR

It has been on the market for two years

Layar is starting to think about standardizing

open to all discussions, not going to provide tchnical details here

AR will be mass medium, new way of experiencing things around you

Dirk was first introduced to the Web at CERN - physics expt

What makes mass medium? The dissemination of content to millions,

The same emerging mass medium is happening to AR

It is a curated ecosystem, people can see what is available and publishers/consumers are coming in loads

LAyar has been incorporated in Samsung phones

AR is really about media specific content

The layar platforms has more than 2M users, 3500 developers, 900+ layers, 1.6M augmented objects served per day

Missing elements: No standards, no global search for content, no interoperability

One use case: Where is the nearest "pizza restaurant, atm machine etc."?

REquest format: knowing the context

Layar uses HTTP GET for "get POIs"

Custom lis of parameters to tailor response

position: lat,lon, alt, accuracy; filter setttings fo a layer

List of POIs: The AR view is a list of POIs

No ordered structure like a web page (Layar uses JSON array)

Another use case: Twitter Layar

Refreshing AR: data in AR might be changing rapidly

Right now Layar uses HTTP to refresh with variable refresh rates


Keep it easy for content providers

Another use case: How does the erlin wall look like when superimposed today?

3D content: Each poi has associated resources

The 3d object creation tools are needed, representation depending on distance

Placement and pose of the object (position, size, rotation), pose with respect to real world features

In AR how do we place object where they belong (e.g. attach to buildings etc.)

Interaction with POIs is to integrate the AR view with Web

Layar will pull web browser model through to AR view; allows for seamless interaction

The web browser is the best way to present AR to the user

on a web page layar://mylayar/?action=refresh&comepaam=3

the AR browser know how to interpret the former url

Layar stream: search and discovery in AR

AR content is not just another type of web content

- Its placed in real space

Its only relevant at its specific location

Its more difficult to link to other content

Crawling AR content is a challenge - what is the role of catalongs? what is the role of content providers?

Summary: to consider for standardization - List of POIs, request including contenxt info, object format, indexing and searching

Panel: Dirk Groten, Claudio, Alex Philips, Timo, Chris Burman

Claudio: been involved several years in AR for telecom italia

Cladio: Feels that we are coming again to walled-garden approach, AR has its challenges; it is critical that we have a lot of data to represent; seeing a lot of emphasis in Linked Open Data

DKA: Linked data is a more generic label
... Alex, can you speak a little about 3D?

Alex: IBM has a handful of AR related projects; a trickle of requests coming from customers
... Enhancing sports events; remote collaboration enhancement;

in AR what is actually holding it up?

1) standards is holding this up

2) technology is still holding this up

DKA: TImo on X3DOM

Timo: X3D can directly be shown without any plugin
... Fraunhofer institute

doing applications for vertical markets, research the use of AR

DOM as base for AR-Webbrowsers

Intersection of HTML, X3D and Tracking Standard?

Indoors, AR will not work that great

Computer vision still needs to scale to millions

AR browser on an iPhone from x3dom.org

DKA: Chris Burman on pachube

Chris: Pachube open platform to create many-many devices

How does this realte to AR?

All this information that users are publishing using sensors; How do we get th estandardize the combination of geolocated content?

Mechanisms need to be standardized to allow users to share content that is public

DKA wants coffee

key things: linked data, semantic web, pachube fits into this

As data becomes more public (govt, geolocated data), how do we pull in the data and discover teh data?

DERI rep works on semantic web and sensor information

Sensor network information do not consider W3C as the defacto standard

Making the data (creation) is more important than making the data public

Alex: There is a whole range of applications that mobile phones are not the best fit

DKA thanks everybody for the panel session

<phila> DKA: Summarises the previous session

<phila> Highlights Jonathan (ETRI)'s 12 points

<phila> Also makes point about XGs. Social Web XG we're looking at social networks etc.

<phila> The work is not just looking at what W3C can do but also the other standards and activities within the eco system

<phila> A W3C activity that might follow on from this Workshop does not mean W3C doing a landgrab on AR, but is an opportunity to build the ecosystem

<phila> ... we have more to talk about than we can in a day and a half.

<phila> An XG could be a release valve for that

<DKA> Scribe: Dan

What Role for Standards

Phil: first session was a firehose of information. In the 2nd session we will look at what role can be played by standards ...
... four speakers to go.
... first of all Klas from SonyEricsson. [slides]


Klas: my area is researching interaction technologies...
... we want to see engagement from W3C in this area.
... "beyond the keyhole" is the title - we are talking about the current AR experience - there is a potential to make this experience better.
... the mobile experience today - you point your camera in front of you - the keyhole effect - also the "tourist effect" similar to holding a guide-book. You have to initiate looking at the guidebook.
... We believe in full peripheral view AR.
... What would that change in terms of the user experience?
... it would be more "continuous use"
... "every second"
... "on the move" - jogging on the beach, talking to a friend, etc... at the same time there is the possibility for augmentation.
... when you have a guidebook you can initiate search. When you have a personal guide it can be personalized to your needs / preferences...
... in order to get that experience you need to get away from the experience of checking off "i want restaurants; i want ATMs, etc..."
... relevance of data is important - how do you make it relevant to the person?
... you need a lot of content in order to make something relevant. You need filtering. Lots of content exists already on the Web. We don't want to duplicate this work. We should take the data that already exists.
... We need to anchor data - to location. AR is about where you are.
... you need some way to have machine filtering. This sounds like the semantic web.
... when writing this paper the first title was "augmented reality: the thing that will make the semantic web take off" ...
... anchoring data - you can use a lot of attributes. Geolocation including altitude, etc...
... You have a 2d barcodes as well.
... merging of these types of anchors needs to be done... Book covers, DVDs, computer vision...
... Filter - number of attributes needs to be there...
... Distance, occlusion, filtering, screen real estate, circle of friends (social aspect), AR volume, user interests, mood, agenda, etc...
... on Facebook you have "like" - people can like and dislike posts (e.g. on Engadget) - the use of participation in terms of making clutter go away - through disliking.
... "AR Volume" -- your view could be plastered with commercial messages, you could turn it down or mute to reduce level of intrusiveness.
... Challenges - if you want all this data, how do you do the aggregation?
... in the same way we already talked about Google crawling - if you're standing in one place but you do a wildcard search with the context "where are you" - who does this search, who does this indexing?
... we want a distributed network.
... User privacy also very important.
... user participation... Gowalla and Foursquare are mapping the world. Users are helping to build those platforms.
... AR needs the semantic web / semantic web needs AR. We need this user context - challenge and opportunity.

Chris Burman: It could be liking and disliking objects in the existing (real) environment.

Klas: I think that what you like and dislike will be shared to only a number of people.

Chris Burman: What happens when those decisions become quite negative.

DKA: Maybe if I dislike you, you're occluded by video content that I do like...

Klas: How well things are "liked" is one way of ranking stuff. In the user context, what is "you"?

Phil: [introducing Raul Roman Lopez]

Raul: I will talk about our position paper. [slides]

Raul It's possible to build real complex applications on the web - e.g. facebook...
... AR browsers are only for consumers - no producers...
... Lots of possibilities for content labeling - tourism, education, games
... We need a open standard. Free Libre Open source software - "FLOSS").
... with Open Source, developers could experiment with new applications, it could become a reference implementation for standards, leading to faster AR apps development...
... We could have innovation sooner.
... LibreGeoSocial - it's a mobile social network with an AR UI.
... this allows users to create their own layers / channels and share with friends.
... uses geo and visual patterns like QR codes.
... we have an advanced privacy system which allows access to e.g. attributes of an object.
... not necessary to create an account.
... [can use e.g. Facebook account to access]
... [goes through some scenarios for user tagging]
... [describing how they tag 3-d space]
... LibreGeoSocial - http://www.libregeosocial.org - client for android, server in Python - soon we will have 1.1.
... we're working now on improving the accuracy of computer vision...
... [concludes]

Phil: Questions?
... about people uploading stuff - the bi-directional web - what kind of info should I upload?

Jose Gato-Luis: currently we only ask for a title and description and privacy information when you upload a picure.

Phil: Do you aggregate that?

Jose Gato-Luis: Yes - you have several options - maybe you're only going to see the content from you friends - we are developing a system based on layers - you could be using a layer that is the official layer of (e.g. Barcelona).

scribe: you could be using the layer of your social network... You have many options.

Phil: Thank you.

Phil: Introducing Jens de Smit from SurfNet

paper [slides]

Jens: it's a bit getting AR applications into Web browser standards so we can use web browser as a platform for AR.
... will show what can be done already with Web technology.
... also how to get local device access, local sensor access into web applications, the security issues, etc...
... [goes through some recent history of AR apps]
... most of these are delivered using WWW technology.
... AR browsers use HTTP for transport, use JSON, XML, etc... Some also use Flash and other proprietary tech.
... Some also run as native executable. e.g. Ray-Ban app is windows.
... this needs to change. AR should be available on the Web across a range of devices.
... [showcases Ray-Ban's virtual mirror]
... Hundreds of videos of this on Youtube...
... but it's a windows downloadable app...
... but it's not rocket science to do this app.
... What we need to build this as a webapp - access to the webcam, microphone, ...
... tracking algorithm to track the user's face. Having tracking tech in browsers would be great...
... can we get tracking algorithms in web tech?
... Transport, rendering, 3d, webgl, javascript for control, html and css for interface markup. All we really need is access to the video feed in the browser.
... more also needed - camera, location, direction, device position. We want that info too - some available in HTML5 [geolocation].
... We want access to sensors - gyroscope, thermometer, altimeter, air quality, humidity, you name it?
... all of those sensors can provide valuable information for context.
... do we need a generic device interface?
... security is also very very important when we're going this way. Once you get info from a webcam into a javascript control you could do anything with it -
... possibilities for privacy leaks are enormous.
... some users don't care (e.g. [the] chatroulette [crowd])
... maybe they just don't know rather than "don't care".
... others may decide against usign the technology at all.

These topics will be discussed in detail at the upcoming W3C workshop on privacy of device APIs, FYI: http://www.w3.org/2010/api-privacy-ws/ which I am also co-chairing.

Jens: [summarises]

Also see DEVICE APIs working group: http://www.w3.org/2009/dap/ and Geolocation API working group: http://www.w3.org/2008/geolocation/

<francois> DKA: [mentioning on-going works in Geolocation WG for device orientation, and in DAP]

<francois> ... I can but echo Jens concerns about privacy and security, and am co-chairing a workshop on that very subject in July in London. Please join!

Phil: If you think those working groups are not going as quickly as you want or not covering what you want them to cover then you need to get involved in them - join the mailing list, etc...
... anyone can get involved - join the mailing list join the discussion. that's how it happens.
... introduces Wolfgang Damm from Mobilizy

Wolfgang: I'm going to talk about 2 standards - ARML and KARML competing ideas? [slides]
... ARML is based on KML - KML is too rich so we based it on a subset of KML. openarml.org
... [displays / explains ARML code fragment]
... We have providers and POIs in ARML.
... we have placemarks - POI - connected to a providers with an ID.
... namespaces: we have the AR namespace which we thought should be for all AR browser. Each AR browser should support tags in the AR namespace. Some browser-specific tags that then might move up to the AR namespace.
... some will remain extensions.
... KARML is from Georgia Tech - https://research.cc.gatech.edu/polaris/[something]...
... They used full KML -
... it's focused on visualization.
... comparing 2 approaches.
... ARML is data driven - KARML is more visualization-based.
... ARML has a concept of content providers - we think it's important to have this in the standard.
... also makes content easily identifiable by the user.
... if user has information on where the content is coming from this allows trust to build up to certain providers.

Wolfgang: we think both approaches are relevant - both have KML as an integral part. We could combine these ideas for a [future] standard.
... [concludes]

[question - who is involved in ARML]

Wolfgang: ARML is developed by Mobilizy - just mobilizy and its content providers.

Dirk: On KML - do you know how all the information you can have in KARML could be rendered into AR?

Wolfgang: [not sure]

Phil: Have you or GA Tech taken these extensions to the open geospatial consortium?

Wolfgang: I don't think so...

Damon: There's be interest from the geospatial side of the AR community... but looking at implementations...

Dirk: The approach of using KML is saying that all the content should be put into this one document. My point is: in the AR view you only want to display what people see "through the goggles" but not lots of interactions in that layer. KML is not meant for interaction. I'm interested to see the link to the interactions.
... As layar as I'd like to see that.

Wolfgang: it's possible to use javascript and KML to have that interaction.

<francois> Scribe: francois

DKA: Question is why the KARML guys did not go the open ARML thing.
... One reason might be that it's not a true standards org, and that's where W3C could help perhaps.

Damon: Another point to look at is who is providing the data.
... Companies that are key players in the field may support different standards

<DKA> Damon: We have to look at who is providing the content. If it's just the geospatial industry they maybe KML is OK but if you're looking at other industries we have to support other existing standards.

phila: OK, on to the panel discussion

<DKA> Phil: Given the time - my proposal is that we have presentation from Mohit and Yohan first.

<DKA> Yohan: I am from IRSEEM france - I am not familiar with Web standards. I am a computer science guy. I am a naive user of AR.

<DKA> paper [slides]

<DKA> Yohan: Our position - it might be possible to allow users to tag region in google street view - we can display this tag [in other contexts]

DKA, don't want me to take over scribing?

<DKA> ... idea of our work - we record all initial information from a phone, we can unwrap the 3d sphere from google street view...

<DKA> Scribe: Francois

Yohan: We combine different tags and information about content. We think that using 2D instead of 1D is extremely exciting for advertisers, campaigns, ...

Mohit: [presenting an example starting with an image from Google Street View and a user image, comparing the images to see parts that match]
... [From there, a portion of the image can be extracted and info or an ad can be displayed in it]
... Major concerns are interoperability, multi-person AR (not the same world for everyone), AR for human (not restricted to games), open AR experience.
... It's all about making AR the Web of today.

Phila: We're going to move to demos. And I think that we should do them in parallel.

Vinny: The one thing that I got from the presentation today is about getting the content that's already there, on the Web, into the AR world, and vice versa.
... We have a situation with information overload.
... We need to deal with thousands of answers.
... Several people have mentioned context as a way to filter information. It's good. But it seems, in AR, that it may be limited to geolocation.

Christine: I don't think anyone's limiting itself to geolocation on purpose.

Vinny: For mobile devices, there are many more context info that could be used. I don't know where this information should be processed, for privacy reasons.
... Where are we, who are we with, what are we doing.
... How do we integrate with building location systems for indoor usage.
... Integration with social networks.
... The third thing could use accelerometers (whether the person is walking, in a car, ...)
... is the person in a meeting, etc
... These are other ways to influence the AR results.
... If I'm walking or driving a car, the horizon is not the same.
... not the same distance you can go in half an hour.
... There is some on-going works on device APIs.
... We could need basic things such as the size of the screen.

Klas: It definitely goes in the same direction as what I presented earlier on.

damon: From a different perspective, talking a lot with developer communities. They want to develop Web apps. Some geo tag in HTML5 perhaps.
... not for technical guys, but for everyone.
... AR enhances the real world, so we must make sure that all companies can take advantage of this.
... The Web3D consortium wants to open up to these companies.
... People talk about standards, but it's a very long process. So we need to be realistic about what we need exactly.
... I hope we all walk away with action items, to make sure people are committed.

phila: Thank you very much. There will be action items tomorrow before we close. That's a very important step in this workshop, indeed.
... If you are leaving before the end, please make sure to take an action item before you go :)
... We really don't want this workshop to end with "that was good, on to the next workshop".
... This is one of the times when we're really looking into hearing about people.
... [organizing demos]


phila: Fatima is on stage from ARPA Solutions, one of the commercial companies working in the advertising space

Fatima: [showing a marketing campaign example that involves a specific T-Shirt
... when you put the T-Shirt on, the system recognizes the T-Shirt
... [technical problems]

damon: one question. We tried T-Shirts, but had problem with women T-Shirts because they were distorted for obvious reasons. How did you work around that problem?

Fatima: yes, you need to stretch the T-Shirt a bit in front of the camera
... [switching to viedo demo]
... [3D modeling on the paper board to speed up development using AR]

The Young Generation on the spot

Video demo of the Box (movie where you push a button and win 1 million, except someone dies)

The virtual button appears in the real box in front of the camera.

Was a very successful campaign. It's Flash, no need to install another program.

[Discussion success of campaign that take advantage of AR from real food packages]

Next demo: Augmented Maps

Marco: Too many POI, we had to cluster results
... When POI are too close, they are collapsed into a single one.
... Part of the clustering algorithm would be capable to stack POI in the order that matters to the user.
... Different views to access the data. Details of image, events.
... When you choose an item, you switch to navigation view.
... An arrow tries to guide you to the POI.
... I'd like to be able to develop that navigation system as a Web app, but it's not feasible for the time being.

Jose: [from Libre Software Reasearch]
... [demo of social client application]
... We have typical features of social networks.
... then we have different layers, similar to Layar or Wikitude.
... When we switch to AR, we can tag things in altitude. So we can have the second floor of the building.
... You can tag things, have Google Maps locate your position, and you'll select the object you are tagging to select the distance.
... Problem is having it work indoors.

Demo of a QR code inside of an AR tag.

scribe: The application scans the QR code and applies the information it retrieves for it to render an AR view of the AR tag.

i/Demo of/Chris: Demo of/

Timo: [video demo of virtual world that uses the accelerometer of a mobile device]
... Another video for maintenance. AR shows where parts have to go, what to do, ...
... I also wanted to show an alpha version of an app that uses WebKit augmented with access to the camera.
... When I start it, I get the camera view, and I also have a Web page that I can interact with.
... The whole thing runs with Javascript.
... The tracking framework is very flexible.
... The 3D object uses X3D and OpenGL to be rendered.

DKA: So if we had access to the camera, and 3D rendering in the browser, you wouldn't need an app

Timo: exactly.

Phila: Final word on Acrossair, our sponsor for tonight.
... Here is a Youtube video I'd like to show you.
... OK, it's been a long day, time to stop till tomorrow!

[End of day 1]

Minutes formatted by David Booth's scribe.perl version 1.135 (CVS log)
$Date: 2010/06/24 09:02:28 $