Wednesday, August 04, 2004

SA2 Project Lessons

Strong Angel II, a research project largely funded by the Pentagon's Defense Advanced Research Projects Agency, or DARPA, has taken place in Kona, Hawaii, in July. The goal of this humanitarian project was to experiment how technology could enable military and civilian disaster-relief people to deal more efficiently with each other -- and with the people who need assistance -- in the turmoil that follows catastrophes.

Dan Gillmor writes about the experience in this article. Various technologies were put to test (62 trials were reported) in austere conditions, ranging from advanced wireless networks, high-quality video-conferencing software, real-time translation software, to leading edge collaboration and synchronization tools. Nevertheless, as Dan points out, SA was about many more things than only technology:

I observed much more than some brilliant technology, I saw how people with poles-apart political perspectives could blend, in common purpose, to achieve remarkable results.

The endeavor was subtitled, ``Designing the Edge'' -- a recognition that increasingly ubiquitous data networks have turned some traditional notions of command and control almost inside out. Now, when people at the edge of networks can get the information they need in a timely way, and get what they know to others, they can work faster and more efficiently.

Experience from Iraq has taught that media awareness and rapid understanding and communication between troops and other humanitarian field forces is crucial; unfortunately, media in the battle field doesn't speak English. Thus, translation software has been extensively tested on extreme conditions (+100 degrees F). As Cmdr Rasmussen, SA2 manager, recognizes in this Wired article, "the influence of the press is enormous and we often have a genuine impediment to understanding the population around us if we don't keep track of what the local media is saying".
In one demonstration, a satellite dish outside a tent was capturing and recording the Al-Manar TV station, a Hezbollah outlet in Lebanon. Audio was extracted from the news broadcasts and converted to text in a speech-to-text program. Then the Arabic text was translated, also by a machine, into English. The results, twice removed from what the announcers said, were approximations. But they captured the gist of the reports.

Among other tools tested, Groove Virtual Office software was used for securely sharing files and communication among all soldiers and sailors, doctors and relief workers, technologists and managers involved in the project.

Laptops were everywhere, of course, and most were running Groove Networks'
powerful collaboration software, which lets people share information smoothly and securely. With Groove, people can work offline and then synchronize data when they connect, in a way that ultimately lets everyone on the network -- without any centralized ``server'' computer -- update to the latest information.

What I found very interesting is how Groove was also used to support (virtually) real-time translation; that is, human-assisted machine translation. Since the quality of the machine-generated translation was not accurate enough, soldiers would add the snaps of the trasncript in MP3 format into a Groove synchronized folder (GFS) so that a colleague/"trusted translator" sitting in the central office in DC could instantly and accurately translate the most important and relevant chunks of information.

This blending of human and machine translation capabilities makes the best use of both. The machines get us part of the way. Humans capture more subtlety, but machines can winnow out a lot of the dross first.

Another interesting observation is that, ever since the WTC 11S attacks , it is evident that centralized systems are a far easier target for the enemy, as Michael Friedlich recalls them: "hard targets".

Firefighters couldn't talk to local law enforcement and phone lines to the world went down in Washington. Yet the decentralized technology frameworks survived and aided in the response.....Only the edge can truly survive.

SA2 was about commercial-off-the-shelf (COTS) products being aimed at planes,
sunk under ten feet of water, dragged through lava fields, and just plain working in a jungle base camp when all that "ran" was a generator and a humvee that came through and picked up the "data" because it served as a mobile switching center. This was a test of extreme computing at the edge where any hopes of a server was a luxury at best, and a myth in a worst case scenario (total devastation). Adaptive processes, culturally neutral workspaces, and decentralized technologies delivered the goods in a big way.


Post a Comment

<< Home