The process of cataloging all of the physical and public things in a city is a mammoth task. Identifying and geotagging all lampposts, grit bins, bus stops, drop curbs, litter bins, telephone junction boxes, parking meters, etc. is something that seems to be insurmountable. Not only that but who should do it?
Obviously the local council has a responsibility, but so do many other local and national organisations like telephone companies, gas and electricity providers, car park owners, universities, oh the list goes on.
But the impetus to start such a catalog seems to be stifled by the thought that it needs to be complete to be useful. I want to challenge that assumption and to suggest a way in which the task of cataloging could be begun independently of the owners of the physical assets.
The idea is to create hundreds, no thousands of QR code stickers, each with a unique URL encoded on it. These URLs would be a unique but random short code, just like those used by URL shortening services such as bit.ly. The stickers should be waterproof, not easy to peel off and contain little other information (perhaps just the URL to allow access by people without QR code devices).
These stickers could then attached to any item in the real world and the URL attributed to the physical object and its location, type, photographs, condition, owner can all be added to the object's record.
But the sticking and the attributing don't have to be done at the same time. Stickers can be stuck and the attribution could be done by the first person to visit the URL by capturing its QR code. "This item hasn't been cataloged yet, can you tell us what and where it is?"
Of course it may be that more than one sticker has been stuck to the object. No problem, the system's algorithms should be able to recognise the difference between two stickers on a lamp post and sticker on the litter bin attached to the lamp post.
Once items are catalogued, the URL becomes an interface for interactions:
"This is a lamp post on South Road. Do you want to:
- Upload a photograph of it,
- Report it as faulty,
- Make a comment about it,
- Claim ownership of it."
By acting as a brokerage for interacting with public physical objects, the project could:
- provide a common set of open APIs for developers
- amass a collection of open data
- show the existence of a digital city in the physical one
- provide a single, unified implementation process in order to change peoples’ behaviour from consumer of public objects to being involved in their maintenance.
Acting as a brokerage would also mean that existing systems could be utilised, FixMyStreet for fault reports (that are a council's responsibility), Gowala or FourSquare for comments or review, flickr for photographs. Using these kinds of services would greatly reduce the effort needed to build a common point of contact for them. The project starts to feel like a link shortener for public physical objects.
Hmmmm, this is just my initial brain dump. All your thoughts are very welcome.
Martin Brook
February 18 2011 - 13:03PM
An interesting idea, this. I had a few thoughts which I thought I'd share...
There's no reason for the basic infrastructure (i.e. the URI->object metadata mapping database) to be limited to a specific city - that's imposing a fairly arbitrary limit on the use of the system. Better would be to create a global, open service and let different organisations in different parts of the world use it in different ways. The service could be set up with template 'object types' - a lamp post has basically the same attributes anywhere in the world - but also allow for new categories of object to be defined as needed.
Also, I wouldn't focus too much on QR codes as the basic identifier - I'd consider the URI to be the unique id for each object. QR codes are a good way of representing this at the moment, given that smartphones can read them, but there's no reason why other representations couldn't be used where appropriate. RFID might work better in many situations, for example. And QR codes will most likely be superseded by something better at some point.
But the other issue that came to mind is that the public objects in question already have a unique identifier: their physical location, which is unique (at high enough resolution) and unchanging (for the most part). Indeed, OpenStreetMap's database already includes some of these types of objects - post boxes and bus stops, for example - and allows arbitrary metadata to be associated with them. Is it necessary to create a new unique identifier for an object when a location-aware device (or a person with a map) could already identify it using existing open data? (Obviously a problem arises when objects are located in close proximity to each other, but this wouldn't be insurmountable because nearby objects - e.g. a litter bin at a bus stop - are likely to be easily distinguished in other ways).
Sorry for the long and meandering comment. (It does say 'brain dump' up there after all).
Saul
February 18 2011 - 13:05PM
Thanks Martin,
You are absolutely right - the QR code bit is just a bait and hook mechanism - something a bit intriguing for people with smartphone to play with. The unique identifier is the important bit.
I'm also all for using existing platforms for particular services. Like 4Square for checking in and commenting, Flickr for photos and Open Street Map for geo-locating. The important thing is to aggregate all of these resources to establish a behaviour for users (citizens) so they know exactly what to do when they want to interrogate or interact with an object in the public space.
Saul
February 18 2011 - 13:07PM
Ah! http://www.digitalurban.org/2011/02/norway-4000-bus-stop-that-tweet-record.html
QRcodes on bus stops in Norway.
Chris Dymond
February 18 2011 - 15:08PM
I'm not convinced by the use of QR code stickers for this purpose either - at a certain point they start impacting the urban environment as much as 'tagging' does, and in many people's eyes they'd simply be litter. Plus I reckon they'd be very succeptable to people 'gaming the system' with a sharpie marker...
Location tracking would work just fine even at current resolutions - I can well imagine choosing from a foursquare-style list of objects in the immediate vicinity...
Kris
March 28 2011 - 16:54PM
I'm not sure. I don't think a location tracking is all it's cracked up to be - I'm forever finding that someone has added something to Foursquare when they have had a bad GPS signal, the result is the object ends up being tagged 100s of meters away. Like my house which is 264m away when I'm standing at the door, maybe this is just a bad implementation by Foursquare tho and averaging geo-tags on a per 'tagger' basis might be a much better way to handle this level of error.