I've taken some time yesterday to further organize my thoughts from this post. We needed such an application on 9/11, we need one for Katrina and no doubt as the hurricane seasons get worse and worse and the potential for other natural disasters like blizzards and earthquakes, we need to create such an applications.
I passed some of these ideas onto MoveOn.org, perhaps they'll take the lead but they're a political organization. The US Goverment has already proven that they're to mired in politics and turf wars to execute a successful aid operation and they're always behind in technology. NGOs and other aid organizations don't really have the resources to create such a system so it's up to the technology community.
Not only is this an app that could be used in the US but with open mapping data initiatives springing up around the globe, and platforms like Google Earth to make use of them, this could be used for any number of worldwide natural disasters.
<Open BrainStream>
1) Trust Issues - It may be too late to implement this in a fast matter but between Katrina and 9/11 it may be wise for a massive organized effort to support such a thing - take the "friend of" concept of Orkut or LiveJournal which gives a level of trust to the friend-of-a-friend level. You see, how many people really want to open their home to a stranger with absolutely no context to who they are? Far fewer than if there was some sort of trust system in place.
I've been builging up my contact list on LinkedIn as I personally see it as one of the most successful applications for Social Networks to date. There's about 5 people I haven't directly interacted with online or in person who are on my contact list but all the rest I have some degree of trust with them. So take 78 contacts with their total 5,900+ contacts, and those contacts' contacts equals 367,400+ according to LinkedIn.
Combine that with Orkut's "levels" of trust: friend but 1) haven't met, 2) acquaintence, 3) friend 4) good friend, 5) best friend. I'd almost add a similar level for family: 1) Immediate family 2) Sibling's family 3) Parent's family etc. Honestly, I'd like to see a flexible filter-like system of LiveJournal where perhaps we have a large set of defaults, (like above), add a few more relationship types and allow people to create their own and set trust levels that are translated into "English" for them to verify.
2) Ridesharing - Both due to a friend's need to get a relative from TX to BOS and since many of the homes appearing on hurricanehousing.org are way further out than 100 miles, people either need to get to friends or family or even these hurricanehousing sites.
I see the ridesharing almost like bus/train connections. W can get X to point A and Y can take X to point B and Z can take X to their final destination at point C. Almost like a massive socialist communing network ala Zipcar.
Interesting....who needs gas and oil? Set up rideshares across the country! ;)
3) Google Maps/Earth mashups - Take all the basic location data for both rideshares and hurricanehousing and stick it in a google maps setup for people to see and have access to. This will be very helpful when trying to determine how to get from one Rideshare location to another. Perhaps hack the Google Directions to say ....."to get to Boston, take rideshare A555 from Baton Rouge to rideshare D453 in Kentucky. Take rideshare D453 to Pennsylvania rideshare T887. Take T887 to Boston's South Station.....or whatever.
<Close BrainStream>
We have the tools - I passed these ideas on to the rdfweb group instrumental in developing FOAF and the Google Maps ideas to a geowanking list.....we just need to reach more geek types. They're out there and we can do this.