I was looking at the NetStumber national access point map and noticing that it hasn't been updated in a while. Clearly the excitement of finding WiFi access points has worn off. But it started me thinking. What if there was a way to use that kind of data--the incomplete, potentially erroneous, volunteed-created--data, because there seems to be something interesting about knowing the GPS location of a given access point? I mean, access points don't tend to move and are broadcasting their IDs (in the form of their Ethernet addresses) all the time, regardless of whether you can log on to use the Net. How could that be useful?
It's useful in this way: until GPS hardware gets to be cheap enough to incorporate into everything as WiFi and Bluetooth are now, using Access Point/GPS location databases can create a cheap kind of GPS for WiFi devices. All that has to be done is that a table matching one with the other has to be downloaded to a device, and then when that device sees the Ethernet address appear, it knows (with some rough degree of certainty) where it is. Imagine a busy downtown area, such as the Financial District in San Francisco. There are dozens of visible WiFi access points per block. Sure you can only log into maybe 10% of them, but who cares if all you're looking for is their Ethernet address? Downtown SF is full of history: this giant area here was once the bay, here was where the Niantic Hotel used to stand, here's where that fire raged in 1906, etc. I can imagine creating a tour of downtown San Francisco that would fit on a WiFi-enabled PDA--something that's likely a whole lot more common than a GPS-enabled PDA--and would allow you to walk around downtown, giving you information for which the location was determined by the WiFi access points visible to the PDA.
Would it be super accurate? Of course not. But it would be cheap from a hardware standpoint--it would be expensive from a data-collection standpoint, but not ALL that expensive. One afternoon walking with NetStumbler would collect all of the necessary data, and it could be updated every six months, with the tour cues synched appropriately.
Sure it's a silly stopgap measure until actual geolocation becomes available, but I think that there's a whole lot more opportunity here than just downtown tours.
Posted by mikek at January 28, 2004 11:59 PM | TrackBackHeh. Too right. It suffers from some of the problems of GPS tho' - wi-fi is notably power hungry. Also accuracy isn't as good, and trust (of the database, and of whoever took the reading) is important.
Same method could apply to cell phones - the cell-id of whatever transmitter you're connected to could locate you. Similar connection method as well (I still need to gerryrig a a GPS to a phone, and write enough software to report back cell id, GPS co-ords, and maybe even signal strength).
I'm talking about location finding at Etcon - will be publishing my talk notes afterwards. You've discovered 1 of the 35 ways in my presentation!
Posted by: Chris at January 29, 2004 12:11 PMJust a note: the netstumbler map has not been updated in a long time because handling that much data was getting to be way too much of a pain.
However, if you slide on over to www.wigle.net you'll see that someone else figured out a more efficient way of handling the data. We are just about at 800,000 AP's.
Posted by: RenderMan at March 19, 2004 02:48 PMThat's an interesting idea. However, the owner of the access point would have to enter that in, and unless they had a GPS, they wouldn't be likely to do so or could put in fake data. I would trust a database collected by a trusted third party.
Posted by: Mike at October 16, 2004 11:16 PMHello folks nice blog youre running
Posted by: lolita at January 19, 2005 04:37 PM