You're all over thinking this (was: Re: Vonage Selects TCS For VoIP E911 Service)

Wed Jul 20 23:19:35 UTC 2005

I don't know all that much about commercial VOIP service or GPS, but it 
seems to me I've just read lots and lots of messages citing weird cases 
where locating a VOIP phone won't work well as evidence that the whole 
idea is a failure, while none of those cases appear to have much to do 
with the problem that people have been trying to solve.  The end result of 
this is that a bunch of people who have loudly written the problem off as 
impossible then start loudly complaining that those working on the problem 
didn't ask them how to do it.

The basic problem, if I understand correctly, is this:  For the last 
several years, anybody picking up phone installed in a reasonably standard 
way and calling 911 could expect that if weren't able to explain where 
they were, the police would show up anyway.  It was hard to see this as 
espionage or as a civil liberties violation -- the wire goes where the 
wire goes.

Now we've got competition among providers of wire line residential phone 
service, and the competitors are mostly VOIP companies who provide their 
service over the users' cable modems.  Since this service is being 
marketed as equivalent to regular home phone service, and used that way by 
lots of its customers, it seems reasonable to expect that calling 911 from 
it would work the same way.  There's a minor problem -- the VOIP carrier 
often doesn't provide the wire, and thus doesn't know where the wire goes 
-- but that seems easy enough to get around.  The simplest way to do it 
would be to ask two questions when the service gets installed:  Is it 
going to be used in a fixed location, and if so, where?  Asking the same 
questions again whenever the billing address changes should keep this 
reasonably up to date.

There are, of course, other ways to do this, which might also work. 
Whether GPS in the ATA box will work has already been discussed to death 
here.  Requiring the cable or DSL providers to map IP addresses to 
installed locations would presumably also work, although with many more 
layers of complexity to go through to have useful information accompany a 
phone call.  Anyhow, I'm sure if we leave those questions to those who 
have to implement it, they'll figure out something that doesn't require 
too much completely extraneous work on their parts.

There are, of course, VOIP installations where this won't work.  I use a 
VOIP soft phone through a gateway in San Francisco to call the US from 
countries where using my US cell phone is expensive, and there are plenty 
of other people who use VOIP phones in much the same way.  Owen maybe 
isn't quite unique in his bizarre scenario of trying to hide his location 
by using his wi-fi phone via repeaters from two counties away from the 
base station.  But these scenarios aren't at all relevant to the problem 
at hand.  If I need urgent help in a hotel room in a foreign country, I'll 
grab the hotel phone and call somebody local rather than trying to patch a 
call through to the US via my computer.  And if Owen were to die because 
he deliberately hid his location when calling 911 and the ambulance 
couldn't find him, it would be hard to argue that it would be anybody's 
fault but Owen's.

At some point it makes sense to solve the problems you can solve, rather 
than inventing new ones.

Yes, this ignores the cell phone issue, which seems rather different 
because they're almost always portable.  It's already had years of work 
put into it, and doesn't need to be reinvented here.

-Steve