SRI has developed a scheme for mobile phones to be able to drop back to peer-to-peer networking if they can't reach a tower. They'd hop over the p2p net until they could get to an uncongested tower -- and potentially even just use the p2p net to make local calls without the intervention of the towers at all (which would make the call "free", since there would be no billing to track).

This is a cool idea, but I'm skeptical. First, there is absolutely no incentive for the service providers to offer or support this service. First, it reduces their revenues. Second, even if one provider broke ranks and offered this, they'd have a hell of a time selling the "increased reliability" story, because the increased reliability depends on some very invisible factors: how many phones are near you, and what the ambient traffic is on them (don't kid yourself; these guys will get oversaturated too). It would be more likely to work where there are denser populations of people -- but that's where there is already denser tower infrastructure. People wouldn't "get it", and so they wouldn't recommend it to their friends -- and the guy who broke ranks to offer it would get creamed. (Kind of like WAP, which offers functionality that behaves so differently than the web that nobody cares to use it.)

And why do the service providers need to support it, you ask? Can't we just get it into phones and route around? Well, sure, except for the part where your call is routed to handset X, and X wants to route it to Sprint PCS -- who gets billed for that?

On the other hand, it would be very cool if you could buy a cell phone base station that lived in your home, so that when you were home your local calls could be routed to your handset, and your outgoing calls on your handset would go through your local service. Let's abolish this crazy world where there's more than one phone number per person.

(OK, OK, maybe three phone numbers: professional, personal, and fax. But can't we ditch fax for e-mailed digital imagery?)

Note that I'm not dissing the whole p2p aspect; I think that's very cool. I just don't think the human factors aspects work. I think this kind of stuff is more interesting for "found environments" where a whole bunch of people come together in a place that doesn't have an infrastructure (like a conference or a cruise). There was some nice research into ad-hoc networking over WaveLAN (the predecessor to 802.11b) going on at Carnegie Mellon (the Monarch project, now at Rice); there's also the zero conf networking stuff ("Rendezvous") that is going into Jaguar. Being able to use computers collaboratively without being in your infrastructure zone is very exciting, and surely there must be a way that bleeds into cell phones that makes more sense that trying to avoid congestion.

For example, here's one more level of CRAZY SHIT. Let's keep the p2p routing scheme above, and tag into it a way to have a "buddy list" on your phone. So if your buddy is nearby, she shows up, and you can say "yo, wazzup, let's go get some lunch!" Because that's an easier story to tell: you can use the crazy new functionality if your buddy shows up. And giving it a slightly different style of working (select from buddy list) gives a cue to the end user that this isn't going through the cell phone infrastructure. (See also ActiveCampus at UCSD.)  12:02:35 PM  (comments [])