« Chicago crime comment spree | Design resources »

June 29, 2005

At Where 2.0

Some very sketchy notes follow in the extended entry. My more coherent take on Where 2.0 will come later, probably after I recover from the 4th of July.

Tim O'Reilly keynote

Web apps all about "what"

But there’s also a "who" layer – databases of identity
When – EVDB
How: Google…
Where: MapQuest, Maps, Local
[Liz asks, "but what about the WHY?"]

Web 2.0 – cooperating data services
- return, reuse, recycle
- patterns being set for how webapps cooperate

1st Gen
- ESRI, MapQuest, MapPoint
- APIs w/usage, but not hacker friendly

Ex: fundrace
- map, neighbor search “very cool”

Google sightseeing
Impact of clear cutting in Vancouver

My childhood, seen through Google Maps (Flickr + Google)
Google Maps plus Craigslist
“this is cool”

Gmaptrack – live traffic cams in London

What are people really doing?

Salesforce.com integrates Google maps

Then Google Maps plus Yahoo Traffic

Yahoo Maps Open Publishing API

Hacker activity teaches industry where to go – engaging with early adopters

[as always, they think that the difference between the desires of early and late adopters is one of size, not kind]

Designing "remixable" web apps
- content as xml, behavior as script
- rewritable web (greasemonkey)

[death of the traditional web designer as primarily a graphic art]

Software is a service; you’re changing it all the time

Users add value (need to bring in tools to help non-alpha geeks)
Network effects by default (auto geodata capture in devices)
Small pieces loosely joined (syndicated business models)
Software above the level of a single device ("all the interesting applications span from the handheld up to the Internet, perhaps with the PC as a control station in between")
Data is the next Intel Inside (data suppliers could choke off industry – Intel did really well: embraced Moore’s law, “the future will be bigger than the past. We have to encourage the ecosystem to grow, to create more demand for the services we can provide. We have to think the same about Web 2.0."…give hackers free access to API calls. "We have to give players room to let innovations roam."

"an invention has to make sense in the world in which it is finished, not the world in which it started" – ray kurtzweil

Connected devices, connected data [where’s the people?]

Google Earth

Inspired by Neal Stephenson [one of many projects - I want to an analysis of the comparative influence of various sf books on computation]

"an interface to the planet"

"my places" – save searches and annotations

"geography as a filter" but also history, news…but also looking for a place to eat

To be useful, information has to be seen in a geographic context – a "substrate" of geographic data

Will release xml schema for describing place that Google uses internally

Global product – not about 20 cities in US, not about US and Canada, about world

No place you can’t drop down and explore

Example: planning trip to Minneapolis
- saving airport
- turning on layers like, lodging
- 3D views
- Automatic mapping vcards
- Constrain search to viewport

mapping panel: A sustainable business for data

Google…was checking email.

MSN is committed to rhetoric! Very committed! Koolaid tastes great!
- they’re doing their own geotagging through crawling
- people wanted white page/yellow page info
- talked FOREVER
- making search a platform

Yahoo Local/Maps
- maps – way of visualizing search
- my reading: we have lots of groups doing lots of things, and we’re trying to figure out how to combine them…[sooner rather than too late, I would hope]

Google Classified Search
-looking at how classifieds work as inspiration

Local Matters
-directory assistance
-ontologists and consumer marketers and software engineers
-for deep structural search?

-"consumer" vs enterprise
-They focus on enterprise and would like to reach consumers

Schuyler Erle – very very enthusiastic about community outreach – seemed like no one was thinking of that - his is a very different version of "sustainable"


John Paul Rademacher
Housingmaps.com – click and see pictures!
Craigslist + Google maps
- people liked it because they like both sites, and he stayed true to the guiding principles of both sites
- maps as platforms – build an entire application around it

Greg Sadetsky
- traffic maps
- broken- api changed, it doesn’t work -- which is a continuing problem

Open Source Geo Tools

- geospatial means money
- bringing hackers and oldschool GIS people together
- GIS manager for Timberline, B.C. – forestry consulting
- Infrastructure
-- Data is central – where to get it, how to use it
-- But assume you’ve got the data
-- Data components in open source setups
-- Databases (postgresql, postgis, mysql)
-- Files (shape files, geotiff, gml/xml, csv/test)
-- Real-time apps (cool factor – real-time GPS, web services, search)
- tied together by Geospatial Data Abstraction Layer (GDAL – data formats)
- functional components of applications
-- visualization
-- manipulation
-- analysis (he’s most interested in this, but newbies are wowed by visualization)
-- creation
-- conversion
-- data access

Lightning Cool Tech

Colin Bulthaup – Squid Labs
- ex MIT Media Lab convened to do cool things
- sensors for unmanned aerial vehicles – UAVS – talk to each other and use GPS to coordinate flight
- self-discovering mesh networks
-location information goes into other things besides cell phones
-lot of emphasis put on position – where you are – very little emphasis put on knowing where you’re looking
-[this is why o’reilly is interesting – it’s polling such a small number of people]
-MagicWindow: combines GPS with orientation from magnetometer and accelerometer and overlays significant virtual data – peer through the world at all the different cities (“View through the center of the earth”)
--Provide virtual information at any point during the day
--Migrate to a cell phone
--Same level of location information to firefighters inside burning buildings

Moto Development group

-elevator display – as the elevator moves, the three panel displays move
-[new industry…coming soon..so soon – DIM, Mike, ITP]
-[location is a funny word – getting close to “site specific”]
-Nice things – staff had favorite floors to visit

Sean Savage
- project placesite
- wifi cafes
- 8500 cafes offer wifi in us
- 75 independent free cafés in SF
- An opportunity for location-based services
--Technology and part of daily routine
-surveys, interviews and field observations in Seattle, Bay Area with computer users and non-users
-vast difference between online and offline experiences
--internet is placeless, same experience
--unlike offline experience, which is grounded
-what about a localized internet service as a middle ground?
--Place site wifi router + openWRT + WifiDog + their modifications
--Does everything a normal router does
--But also serves a local service
--Nobody has to install any special software
--Who’s here now?, message, discussion forms, who WAS here
--Starts with personal profile – you can hide or show based on location in café
-now working on place site as network – so that any café in the world can set it up, as for apartment complexes
-placesite as platform, so outside developers can hook in without outside services

Nathan Eagle: Modelling complex social systems

-we are about to get inundated with masses of data on human behavior
-inferring ideas about human activities from wearable human computers
-Nokia gave them phones, they put context logging software on it
--Tracks location, and asks people to name location
--Also tracks, using BT, who’s nearby
-who’s hanging out with whom, and where
-gave them to people at MIT
-45000 hours of continuous human behavior
-Use in epidemiology
-Boston: what happens when the Patriots win to the transportation systems?
-Using Bayesian networks to train the computer to classify a situation
-Shannon information entropy
--Estimate of the amount of structure/randomness in a subject’s routine
--Which demographic is the most entropic?
---Structure can be quantified
---Freshman are most entropic, staff are least entropic
-correlates to how well we can predict what people will do next
-our days have inherent structures
-with eigenbehaviors, predict next activities
-goal: create aggregate usage statistics
-also logging phone usage
--what applications are used?
--Where are they used?
--Ie, Media Lab students use snooze more than Sloan business school
-automatic diary generation
--life query
--how much sleep did I get last week? How much time did I spent traveling? When was the last time I had dinner with X? Where did we go next?
--Prediction: will my girlfriend call me tonight? Will I be able to see my girlfriend next week?
--Interface to visualize and tag experience
-inferring relationships
--friendship vs proximity data
--is it possible to infer friendships from contextual data?
--Is possible to infer team relationships?
-inferring organizational rhythms
--Looking for underlying signatures of group deadlines
--Looking for organizational rhythms
--BT boxes that tracking BT devices around them
--Tracking how fast people walk
-making recommendations
--who should meet?
--MetroSpark – introducing people [aka, his start up that he wants funding for]
-no universal equation for social behavior, understanding constructed out of examples
-informed design of organizations, school, cities, office buildings that reflect people’s actual lives
--well, we’re already there with call logs
--it’s about ownership of data
---right now, the government can get your data, but you can’t
-this is the first and only talk where I saw standing applause

Mapquest Find Me
- find yourself, control access
- need to have alternative to gps
- find stuff nearby
Teen Arrive Alive

Have to get application into handset
Product feedback is vital
Difficult to prove yourself to carriers

Mapping panel

Sprint LBS - Paul Distler
-"had it not been for 911, we’d all be waiting much longer" [which seems to be why they're so unenthusiastic about supporting LBS - it wasn't in their business plan in the first place]
-Emergency svcs
-Enhanced directory assistance
-Roadside assistance
-Asset tracking
-Field force mgmnt
-Development pipeline
--Svs that are consumer priced to drive adoption

Location svs market constraints
-is the system REALLY scalable?
-Need several generations of dev to get to a really usable svc
-Needs continual reinvestment
-Apps must operate in the handset environment – smaller and more efficient
-Partners often don’t develop efficiently to use all the capabilities
-Partners need to generate accurate and
-We need better UI experiences
-Must differentiate across customer segments at prices that reflect inherent value

Nextel – Mary Foltz
-wants "small pieces loosely joined" services rather than the "one ring to rule them all" plan [her words - though clearly she's trying to differentiate Nextel from Sprint]
-Verizon: one ring to rule them all
-Open platform: developer.nextel.com
-Growth has been enormous, starting with first GPS phone 2002
-Emph on developers creating their own svcs
--Laundry list of svs enabled by Nextel
--Like offroad navigation, geotagging photos (“photomarking”), GPS Blackberry
--TeleNav, Trimble Outdoors, Xora
--Field service automation
--Delivery and supply chain mgmnt

Orange - Carl Shimer “the hacker guy”
-he designs and architects application
-more like a 3rd party developer
-problems – getting apps on headset, getting access to GPS
-diff between telco and Internet
--people in wireless realize innovation is important, but it’s much more about making money off of it
--Google and Yahoo will deploy services without knowing where the money is coming from,
--uLocate is a similar example in wireless, but a lot of people haven’t been able to succeed
-send telco execs a copy of “the innovator’s dilemma”

[Carl seemed to go over pretty well]

Hassan Wahla – TeleNav
-application developer
-developer programs w varying degrees of supper
-different operating systems – memory management issues
-handset deployment
--manufacturers, screen size, gps performance, memory
-wireless network LBS implementation
--and then trying to work with different carriers
--mobile station based vs mobile station assisted
--network based location only
-often not building new product – just trying to keep up with changes to the handsets they support
-having applications billed by carrier has aided adoption – avoiding customer getting two separate bills
-bundled data and service better than separate fees
-monthly subscriptions vs transaction fees – transaction fees can help people try the service out
--over the air download is okay for business
--but preloaded software really helps drive customer adoption
-working with carrier to create awareness within their sales force: commissioning and training
-awareness for customers: outbound marketing


Hassan: easy to get SDKs, hard to get publicity and word out

Mary – We never get enough beta handsets before they’re released. Access to handsets is key, but often developers have to wait until the handset is out on the market. We’re a small team and overloaded, so we try to make it easy to at least get started – get going and post applications. If they can get their apps out, they can go to Mary’s team and show them what successes they’ve had. Then they sit down with Mary’s team and do a tech and business review. Because it’s scary to take an app from a small startup and offer it to a large enterprise customer and promise support over the next few years. They ensure that there’s support.

Carl – Orange is based on cell id rather than GPS. Talking about starting in a garage: you have to be incredibly pragmatic about your chances with a carrier. 100 people come through the door, maybe 2 get it. As carriers, we need to do better. There are very small teams in carriers that do LBS. There may be a small market for some of the svcs, but can we try to get them out to the people who will use them? Also, reform fee structure. Ie, Sprint – fees for accessing location-based - move all the fees for application signing and certification.

Sprint guy: Java rather than Brew. Have people who do great demos render a more professional vision. Had great results. But still, only a few make it through the pipeline.

Q: Carriers prefer a walled garden approach. And they’re taking 40% of the (gross?) – taking away the margin.
Beattie: Question is, is there a path to openness?

Sprint guy: The difference that the value chain is integrated. Even though they run a network, the carrier also has to validate and support handsets. The ROI – data does not return. Voice returns. And it subsidizes the data. We

Mary: We’re investing for the long term in the infrastructure. But we need our providers to succeed – we need them to succeed. So we carry some of the costs. We depend on folks, and we count on them to succeed.

Hassan: Praises “tremendous support” from Nextel. The carrier is getting revenue from service fees, but hopefully they’re also getting money from data fees. Hopefully, the customer is more likely to stay with that carrier and get more services. From the dev side, that’s the argument – although the finance side doesn’t always see it that way.

[tremendous annoyance about fees. Sprint is sounding unhelpful, Nextel sounding very helpful]

Q: Server side processing on applications?

Mary: We (and Sprint) have ambitions to do that. Launching API (?) – got it set for max frequency of query ever 10 min.

Stephen Randall: Everything you know about mobile marking is wrong

Connecting bricks, clicks, and pockets

Lazy thinking on both sides of the Atlantic
-"Americans don’t get wireless"
-"Europeans are crazy"

But it's a huge untapped market – left UK and moved to Boston
-halfway between Europe and West Coast

Mobile: because of short term thinking, collaboration has been impossible between carriers and handset manufacturers
War over owning user experience hurts developer focused on innovation
Unless it changes, many of the things we've seen today will not be ubiquitous for four or five years
Consistently, it's five years after the hype when things take hold
Push – only good for Hollywood (and carriers) (ie Minority Report adverts)
Until recently, the carriers believed they owned location – the model that location based services will push advertising to customers based on their location
Which is not what people actually want – they'll turn their phones off
Push is now out of the question – it has to be opt-in to be relevant at all
Maybe the answer isn't to get apps into the device
It's really hard to do – difficult for carriers and handsets
The "volksphone myth" – that all handsets should do everything
One handed use – two handed use stop you from being mobile
He’s fixated on ubiquity – given up thinking about "killer apps" – but it should be useful for hundreds of millions
Three "mindless" clicks (ie, three perceivable actions) – so bringing up a browser is dead in the water

Different view of location
-comes from observing a uniquely American phenomenon
-TV is becoming problematic for advertisers
-Phone is becoming the "third screen"
-But there's a fourth screen – the billboard, which is becoming connected to the internet
-People describe their location quite differently depending on their context – systems are very bad at doing that
-But if they are in close proximity to a large physical screen, you know their location
-If it's connected to a phone, you can do things that are not as compelling on a cell phone
-we operate in safety zone
-cell phone defines a zone of experience – we can pull instead of being pushed
-when inside safety zone, we can foster trust because we’re in control
-here, now and simple (look at Psion pdas – everything immediately available to user)
-instant gratification
-low tech – for now
Pull – signals that show the way
-ie, text messaging to vote on American Idol – look at behavior of kids
-TiVo – inflection point of advertising on TV, driving advertisers away
-Spam – becoming illegal
-Malls – Simon Malls one large mall group in America touches 2 billion consumers a year – once in shopping, now in entertainment business – how do they add value to stores by reaching mobile consumers?
-Walmart – fifth largest TV network is Walmart's
Vision – screens that are already in stores and elevators will go everywhere
Scenario – interactive retail networks
-realtors have screens on their windows that show slideshows of houses. You dial a number on your mobile phone to get more info/pictures about that house
-because realtors can't work all night – so they can be responsive even when they’re closed
Wiffiti – wireless graffiti
-texting messages to screens
-online, you can see lots of different screens from lots of different locations
messages on TVs in bars as well –

panel – privacy and location data

Pamela Samuelson
Kinds of privacy
- Spatial privacy
- Communications privacy
- Information privacy
- methods for privacy
-- opt out
-- notice and consent
-- flushing data
-- licensing terms
-- new legislations
Why respect privacy?
-because it's the right thing to do
-because your reputation depends on it
-because it gives you a competitive edge with customers
-because self-regulation will avert gov't regulation
-because it's going to happen so you should be ahead of the curve
-because international rules are stronger than us rules
So why not design it in to the tech right now?

Panel discussion

-computing devices with a radio and a smart card: it’s an identity tool
-talking about reputation
-starts to look like community that looks like a small town

Q: Is Nathan Eagle's work disturbing?
PA: Not thinking ahead of what might happen down the line with the data.

PA: black boxes in cars. Who has access to that data?

People opt in to an alarming extent – but once you opt in, you lose control.

Q from Police officer: I don’t understand why we have this discussion about privacy. He worked for a department that tracked down the killer of a girl by using his cell phone.

PA: Law enforcement has legitimate interests. But you shouldn’t just be able to get that information at the drop of a hat. It’s ripe for abuse. Need to show legitimate need for that info.

Police officer: So there’s no problem in collecting it, as long as it’s kept in escrow somewhere?

[fascinating – this is the first moment where we talk about the human emotions in location]

Ross – introducing points of failure and risk by keeping gigabytes of data. There needs to be audits and accountability. There are all these pushes for clandestine access to stuff.

Note: I'm not sure which panel this was. I think it was the one called "Retooling Developers for Geospatial"

[the panels became a blur after a while]

The geospatial industry is comfortable with continental drift – but they're dealing with another industry that thinks a year is a long time

MS guy – evolving towards standards

What should we make open source, and what should we keep close to our vest? They’re in transition.

Q: There’s data, and there's representation

Online mapping – the person who has to read the map will actually create the map, and they didn’t go through GIS schooling. It's changing the role of the GIS person. It's no longer to just deliver maps, but to relay knowledge to support the rest of us in understanding how to interact with geospatial knowledge.

Q: Cumulative inaccuracies from many layers of datasets?

Stamen Design: MoveOn's virtual town hall

Moveon virtual town hall
Very large groups of people around the country can participate in shared live events via interactive maps

Map is not reflective but created

Before, used conference call – but the experience was expensive and not all that great

You call in and enter a zip code to be located on a map
You see where you are, and see where other parties are
As the parties progress, you can see how everyone else is doing
There are producers who mediate the feedback and push prompts for action

Borrowing "Kirby dots" to show participation across country – convey energy and excitement

Conclusions: technology
-no “pro” GIS tools involved, just six weeks of open source tools, flash
-the technology is no longer the hard part
-design is the hard part
--embracing uncertainty, not defending against it
--need directed interactivity and considered feedback loops
--design not just of visuals but of experience
--simultaneity: you are part of something
--maps are easily understood: instant recognition, understanding, and plan for action
-moveon used map to ask people what to do AFTER the election
-feelings of membership, influence, needs fulfillment, shared emotional connection

Kevin Slavin: Big Games

Marco Polo – stories about adventures that may or may not be true
-maybe it was so ordinary he had to make things up
-places need stories to make them seem real
-most places are real. Most stories are not.
-Big games are the most real, most fake stories we can tell
-Marco Polo’s stories about Burma opened up trade
-We look to stories to see what else can happen
-Large scale real world multiplayer games
-Area/code: his company with Frank Lantz
-Transform the world around us to a shared game space
-Games with computers in them, rather than the other way around
-People go to places because they have meaning – PE Island and Anne of Green Gables
-Or El Dorado and the Amazon
-Joan Didion: We tell ourselves stories in order to live
Omaha: a big game story
-a 20 ft high big horn sheep
London: as monopoly board
- GPS enabled cab – depending on where the cab goes, you get revenue
Manhattan: filled with pacman-eating ghosts
-a way to reexperience the city on a fictive level
some examples
-commercial, some university based, some art
-some on phone, some without phones at all
-based on game-like dynamics in physical space
Return to Omaha
-Conqwest: 2004 à
-Inspiration: movie Kevin saw when he was 11: “They Live”
-Alien robot droids control the earth, and all the ads are code for something – you can read code with special sunglasses
-Signmark weirdos: the stickers on the backs of signs are part of a secret map of the world
-Semacodes: barcodes recognized by phonecam
-Location from the bottom up: if the phones see a sticker, then we know where the codes are
-dependent on 6 types of services, all of which could and would go down
-the work of getting every city ready – carrying around all the phones and chargers!
Programming the city; dividing it into zones
-controlled through enormous totem animals
-overhead view online and at hq
-but using simple stuff – no fancy phones – all we could do is pull semacode in as mms and push msgs back
-as far as the kids are concerned, it’s all happening on the phone
-low level emergent behavior
--rotc team had subteams
--chain of command
--strange markings on arms
Success is measured by cheating – if people are this invested, then they’re doing something right
Next generation
-more devices that integrate better or different
-more integration, less interruption
-persistent, pervasive, rather than event-focused
-smaller big games
-bigger small games – games on phone should take context into account
-cross-media integration – take TV into account
-rethink location – not just GIS data
-bigger lies better told
Why people play for them
-because TV is good for location-sensing, and bad for advertising
-push innovation, just like on PC
-because progress often starts with play – the ways that we misuse these technologies are the most valuable way we use them
What to do with what we have
-location à dislocation
-information à disinformation
-getting there vs getting lost
-know where I am vs forget where I am
-emulation vs differentiation
-reportage vs storytelling
-accuracy vs misrepresentation
-where 2.0 vs here (beta)

Posted by egoodman at June 29, 2005 3:21 PM


Post a comment

Remember Me?

(you may use HTML tags for style)

Trackback Pings

TrackBack URL for this entry:

Listed below are links to weblogs that reference At Where 2.0:

» confectious: Notes from Where 2.0 from Lifeblog
comprehensive notes. for me of course, the mobile stuff is where my head is at. why can't the operators just get out of the way? they own some of the juiciest location info, apps can really easy integrate automatic location [Read More]

Tracked on July 5, 2005 1:25 PM

» word power and my delicious habbits from Floozy Speak!
Words.  They are powerful.  They are so amazing, so useful, I wish I knew more words.  I wish I knew all the words that would be tags.  Sometimes I surf endlessly thru delicious and wonder what's another /tag/"something" i could... [Read More]

Tracked on July 22, 2005 11:05 PM