Archive for December, 2005

Further Thoughts on Disaster

Further Thoughts on Disaster

I’m not suggesting a Web panacea for humanitarian aid. Disaster is not flickr, not Wikipedia. The issues of accountability, reliability, protection are serious and real. Katrinalist and other disaster wiki projects didn’t experience problems of non-cooperation, but it’s easy to imagine open disaster coordination systems being gamed .. like the stories of some Sri Lankan fisherman receiving new boats, even though they were not affected by the disaster. That’s partially a beauracratic, systematic error .. and underscores the necessity for automated trust, attention and attenuation in open systems (if that sounds like the etech theme, it’s part of a potential session there).

One surprising thing on GDACS, the software is developed in a propreitary-ish model (the code is technically free, but you need to go through an official request process). This is not for lack of desire for open source, but rather institutional inertia encoded in licenses and beauracracy. With the altruistic mandate and requirements for transparency, open source software development by agencies and NGOs seems natural. Interagency projects only stand to gain by adopting an Open Source model of development, a community model more inline with the complexities of coordination among all the current, and potential, stakeholders. There are other potential costs of organizing OS development, but funding the project becomes more flexible, open to whichever agency currently
has resources to contribute. Agencies can interact with the project independently; competing systems can be engaged by shared code.

The common conceptual model and standards of disaster data is not a given. GDACS is self-limiting to “sudden onset natural disasters”. The line here is blurry .. when exactly does a disaster event start? An earthquake is clear. Perhaps also a volcano, though volcanic events are becoming more accurately predictable — does it start when the prediction is 75% certain? The path of hurricanes is somewhat predictable. The speed of a flood flowing downstream is calculatable (though flashfloods are less so). Drought emergencies are complex combination of natural, ongoing processes and deliberate unpreparedness. By these agencies cooperating in a technical manner, they are forced to come to common consensus on definitions.

Mapping in these agencies is another major openness opportunity. The JRC has a Digital Map Archive, UNOSAT gathers and processes satellite data .. all of it somewhat in the public domain, but not so readily accessible. Even though the JRC is coordinating INSPIRE, commercial data is licensed and used, considered to be of more consistent quality than national mapping agencies’ products.

Sounds like a case for OpenStreetMap. They got the usefulness immediately, for many places where these agencies operate have no digital mapping data whatsoever. We discussed how agency field workers could use and contribute to the project. Potentially, they could dynamically create and share maps as they are deployed. One thing to address is the technical constraints in the field, where limited connectivity would require some additional affordances in the OpenStreetMap model: offline editing and conflict resolution would need more support. Even with limited connectivity, something close to real time locative coordination and needs assessment is a huge potential field application. Accuracy is another issue; mislabeled places have resulted in official diplomatic incidents. The UN Cartographic Unit maintains a list of official, treaty agreed, place names. ReliefWeb protects itself with a disclaimer .. “The names shown and the designations used on this map do not imply official endorsement or acceptance by the United Nations.” For OSM, I’d add “You are free to fix it yourself”. 🙂

Comments

Global Disaster Alert and Coordination System

Global Disaster Alert and Coordination System

Last week, as a guest of the JRC, I attended a technical meeting on GDACS, a cooperative effort of several UN agencies and the European Commission to develop a common alert and coordination mechanism for natural disasters. Each agency monitors different types of natural disasters (earthquakes, hurricanes, droughts, floods, volcanoes), and there’s a real push to avoid duplication of effort in disaster response.

The key to integrating data and systems of these orgs is, fantastically, RSS. To me, this use is potentially one of the greatest successes of RSS, and shows how the architecture of the web can be a huge boon to humanitarian and ecological efforts. Surprisingly, the participants had a much more sober view on RSS, seeing the format as just one less thing to worry about, freeing up the system to “fluidly develop capacities”. They are well aware of the latest developments on the web, and consider the methods and technology as needed (Digital Identity and RESTful web services were discussed). The GDACS system makes copious use of namespaces to encode various domain specific information, like the type, severity, and location of the disaster [feed of recent events]. Because it’s RSS, it can be used in other tools, like human new readers, and worldKit (the JRC’s adoption of RSS and geoRSS lead to their use of worldKit, my involvement in the online tsunami model, and tagently the current effort to officially extend the geometries and compatibility of geoRSS).

Perhaps it’s the unintended consequences of RSS that are most interesting. GDACS, GLIDE numbers (a globally unique identifier for disaster events), Virtual OSSOC (a private message board for relief workers), etc. are intended for use of humanitarian relief professionals. But yet, those RSS feeds are out and available for the taking. Any enterprising hacker has access to the same data feed the world’s biggest humanitarian agencies use. The relief orgs are aware of the amazing bottomup disaster responses to the tsunami, katrina, pakistan .. but don’t quite see how it really makes a difference to their own practices .. yet (well some more than others).

There are definite systematic limitations in disaster response. The agencies are beholden to donor nations. They are in competition with each other for the same funding — there is a tension of cooperation and competition even in humanitarian aid. If donor nations do not respond, emergencies can be passed over. To some extent, media exposure and its feedback cycles drive the magnitude of the response. Agencies must be cognizant of the media in order to attract continued funding. These agencies do have resources to draw on, but are plagued by the same inefficiencies and inflexibility of any large beauracracy.

The Web has distributed the power of communication. People can speak directly to each other, and to wide audiencies, without going through media monopolies. This power is needed among the most desperate — to directly communicate their needs, during an emergency and long into the recovery after the media glare fades, and to coordinate directly with those that can help, professional and amateur humans. I don’t think grassroots can do it alone, not even close, but the big orgs need to be engaged, a process of give and take, until the edges of those organizations become as fuzzy as a web api.

Comments

geocoder REST update

geocoder REST interface update

Now US zips and UK postcodes can be looked up via the REST interface. And 3 letter ISO country codes can be used.

Behind the scenes, the script has been massaged to be as efficient as possible. Every microsecond counts, as this popular geocoder is presently running on shared hosting, and using significant CPU.

As always, please get in touch if you are using the REST interface, so I have a contact for any issues that might arise.

Comments

boicentrum.se

boicentrum.se

“Bo i Centrum” is Swedish for “live in town”, not the usual state of things in the far north of Sweden. Skellefteå, my wife’s hometown, is 14 Swedish miles from any other town of significant size (~140 km; distances up here are so great they just divide by 10).

boicentrum.se

This sign (1 of 3 planned in town) advertises the site, outside the company offices. The pic was taken at 2pm, already dark here!

boicentrum.se lists apartments for rent by Älvsbacka Förvaltning, the property management company run by my father-in-law. This is currently the only website around these parts for finding apartments; think HousingMaps for the Arctic Circle.

For this project, I developed a php backend for managing online listings, with integrated mapping ala worldKit. Maintenance requests are also received via the website. Potentially useful for other companies (get in touch if you think so). The imagery was obtained from the state mapping agency, Lantmäteriet, at a pretty reasonable price. It’s been interesting, learning about the byzantine world of Swedish real estate, and making a mark on this town.

Comments

OpenStreetMap

OpenStreetMap

Been getting involved in OpenStreetMap. A great effort, which gets more exciting and interesting the more I think about it.

We recently updated the site with a fresh UI — including ‘industry standard’ features like Javascript slippy map and geocoder.

Today I put together a basic monitoring map of edits in OpenStreetMap. It’s using a new GeoRSS feed of OSM activity.

Comments