Towards an Open Geo Web: Linking Open Source’s ‘Architectures of Participation’ to the Global SDI Initiative

Post on 15-Dec-2014

4601 Views

Category:

Business

1 Downloads

Preview:

Click to see full reader

DESCRIPTION

Talk given at GSDI-10, about bringing the lessons of open source to spatial data infrastructure development.

Transcript

Towards an Open Geo Web: Linking Open Source’s ‘Architectures of Participation’ to the Global SDI Initiative

Chris Holmes

Architectures of Participation

• Term I use for 'the open source process'• Coined by Tim O'Reilly • Strong roots in Yochai Benkler's 'commons-

based peer production' of Coase's Penguin - I take an architecture of participation to be the 'thing' that results from a commons based peer production

• Stephen Weber also hits on main ideas in The Success of Open Source Software

Examples of Architectures of Participation• Linux

– First open software project that truly empowered potential contributors

• Wikipedia– Very low barrier to entry, lets anyone get

involved in the area they know

• Firefox– Brought participation to ‘marketing’ - raising

awareness of the product

• World Wide Web (and web 2.0)

A Definition

• An ‘Architecture of Participation’ is both social and technical, leveraging the skills and energy of users as much as possible to cooperate in building something bigger than any single person or organization could alone.

Social and Technical

• It is not possible to simply focus on one, to take the other as a given

• Community won’t work right if the technical doesn’t support it properly - can’t make it too hard to participate

• Technical can’t do anything if there’s no community - can’t just build and assume people will come

Applying to Geospatial

• Software was only the first domain to see the incredible benefits of succesful architectures of participation

• Steven Weber is clear in his book that the process can be applied to other fields, most any digital good has high potential

• For geospatial there are two main possibilities– Creation of geospatial data– Sharing of geospatial data

Geospatial Data Creation

• Open Street Map is the clear early leader in this domain

• Many interesting issues with licensing, and how to incentivize properly

• Huge potential for cooperation among citizens, governments and commercial data providers

• Unfortunately outside the scope of this talk

Geospatial Data Sharing

• Primary goal of all Spatial Data Infrastructure initiatives is to share geospatial data– To pick one definition: “the sources,

systems, network linkages, standards, and institutional issues involved in delivering spatially related data from many different sources to the widest possible group of potential users at affordable costs” (Groot & McLaughlin 2000)

The Success of SDIs?

• If the goal is making it easy to access geospatial data, then Google in its private infrastructure has achieved far more than hundreds of national initiatives

• 9 million MyMaps vs. ‘thousands’ of NSDI users

• Google is not an SDI, but it’s useful to examine how it’s achieved the goals of SDI so well.

Architectures of Participation: Factors for success• Working on something useful (or at least

has the promise of being useful relatively soon)

• Viewing users as important contributors.– Give them responsibility, see them as the most

valuable resources and they will become it

• Lowering barriers to contributing. – First time contributors won’t come back if it is

difficult to have their effort incorporated to the commons

Google: Contribute to something useful• Very quick win of being able to visualize one’s

data on top of extremely high quality base map and additional layers

• Relatively easy to customize for even more advanced visualizations

• Easy for others to download if uploaded, nice layers get recognized and blogged about

• Gets indexed by GeoWeb Search, others can easily find it.

SDI: Contribute to something useful• Putting data on an SDI seems to have

few clear wins– Few people use SDIs, so you get practically

no feedback• Admittedly a bit of a chicken and egg problem

– Only can show something off if lots of time and money has gone in to making a ‘portal’

– Only can search on the registered catalog – No interesting visualizations

Google: Users as Contributors

• No discrimination between consumers and producers

• Everyone encouraged to upload data, even of things that others won’t care about

• Google has even hired users who have contributed extensive data, to become paid contributors

• Many ‘experts’ have also contributed very valuable data, it’s much easier than getting on an SDI

SDI: Users as contributors

• In SDI’s there appears to be a fairly clear line between producers and consumers

• Data generally comes from national agencies and ‘official’ sources

• Users get the (correct) impression that training is needed to fill out metadata

• Only ‘GIS professionals’, with certificates and degrees are worthy of contributing

Google: Barriers to Entry

• To share data on Google’s infrastructure you need:– A browser

• Google MyMaps lets you create new geospatial data directly with your browser

• Predating that Google Earth lets you save any data you make as a KML file, which they then have a site that you can upload it to

SDI: Barriers to Entry

• To share data in an SDI you need:– Metadata: filled out by someone with

appropriate training– A server to put your data on– Software implementing WMS

WFS/WCS standards– Perhaps data sharing agreements– Registration of metadata on a catalog

But does user contribution alone make an SDI?• No… Not quite• Commercial players have stayed out of

encouraging creation of ‘real’ GIS layers– Though many ‘experts’ exploit commercial

platforms as great visualization tools

• They use ‘real’ GIS as a base layer, but they keep it private

• Not in their incentive structure to make the public GIS data they gather available to all.

Let commercial players run SDI?• Though they’d love that this is an awful idea

– SDI’s are a public good– Commercial players all have a profit motive– Each would like to own the entire infrastructure, to

get to a monopoly where they control the infrastructure and can extract whatever they want out

– This is a real danger, many governments are handing over data without opening it to anyone else

• Only those who have the resources to have workers constantly try to extract data will have it

• Turns public good in to opportunity for private gain

Towards the Geo Web

• GSDI should be a single, shared infrastructure that all participate in.

• Term ‘Geo Web’ emphasizes a single project, instead of different initiatives

• Has the potential to go much further than even the full infrastructures that Google and Microsoft are building

• Build on what’s out there, on successful architectures of participation

Principles: Towards an Open Geo Web• Don’t think of SDI as just a set of data sharing

policies, top down requirements and mandates– At its most successful it’ll only cover government

agencies and a few more

• Figure out how to set up architectures that aligns the various incentives of all those who create and consume geospatial information to participate in a single geospatial web– Resist just letting big companies run it, as it will lose

its key property - that it is a public good.

GeoPortal 2.0

• Vision of an SDI/GeoWeb node that is built to encourage participation in a number ways

• Most ‘portals’ died with the dot com crash, web 2.0 puts the user first– Flickr, facebook, delicious, youtube, ect.

• Provide real value to users, reward them for using and contributing

• Bring participation in to every aspect

GeoPortal 2.0: Evaluating datasets• Provide statistics and rankings on how many

times layers are accessed • Enable commenting, rating, tagging and

reviews by users• For providers that allow it, wiki style editing

of ‘official’ metadata records• Amazon style collaborative filtering based on

user profile (rated high by city planners) and similar users– Rating and tagging more gives you better results

GeoPortal 2.0: Viewing data

• Easy to add a bunch of layers and create a ‘view’

• Save your favorite views• Link to ‘embed’ on blog or webpage• Rate and comment on views• Others can start with another user view

and remix ( amd evem style and add data)

GeoPortal 2.0: lightweight data contribution• Easily add personal annotations to a view• Easy to start a new data layer, that others

can collaborate on– Set options for sharing, wiki-style anyone can

edit, only approved users can edit, ect.

• Enable easy online editing, but also editing through GIS with WFS-T

• Simple upload of KML created from Google Earth or GPS traces

GeoPortal 2.0: GIS data contributions• Simple form to upload a Shapefile • No required metadata, but prompts for

some easy fields to fill out, or upload of FGDC/ISO form

• Set up permissions on who can view and edit

• Easy online styling of data for default view, or import ArcGIS style files, SLD, ect.

GeoPortal 2.0: Data availability• All data contributed becomes available in a

wide variety of formats– WMS, WFS/WCS– GeoRSS, KML– Cached tiled rasters, Shapefile

• Can be used on Desktop GIS, web-based slippy maps, on Google Earth and Maps, Virtual Earth, easy embeds for blogs and webpages, ect.

• Crawlable by Google’s GeoWeb Search

GeoPortal 2.0: Registering Services• Easy to register a new geographic

service (WMS and WFS/WCS)• Registered layers gets tiled and cached• Tiles are available on Google Earth and

Google Maps• If back-end service goes down cache

lives on• Can limit caching to ‘official’ services, or

just have it cache the most accessed tiles

GeoPortal 2.0: Metadata

• No metadata requirement to use the infrastructure

• Derive metadata from user’s active and passive actions

• Enable wiki style editing of metadata records

• Make metadata creation a ‘game’: recognize users who contribute the most good metadata (both editing and ratings, comments, reviews, ect.)

Official vs. user contributed data• Providers of official data should use exact

same infrastructure as ‘amateur’ users• Can limit permissions to a controlled area

before official ‘publication’– At publication time metadata record should

be complete, data verified, ect.– Before it can be internal wiki, editing, ect.

• Like on youtube can disable comments and ratings

GeoPortal 2.0: Technology

• Future is not about SOA - it’s about users• Enable participation all the way down,

software must be open source, standards used must be open, should be seeded with as much valuable open data as possible.

• Technology and community are always linked, GeoPortal 2.0 must be built iteratively and be useful every step of the way

My SDI/GeoWeb Goal

• Internet made it so citizens demanded e-government, let’s build a GeoWeb that’s so compelling and easy to use that the question ‘why isn’t my government making its geospatial data available’ comes from everyone, not just ‘the experts’.

Learn more

• Justin Deoliveira’s talk on GeoServer at 15:00 today room 1

• Stop by our booth

• http://geoserver.org

• http://opengeo.org

• http://cholmes.wordpress.com

top related