Show HN: TT911 – Text 911 your location and name without unlocking your phone | Hacker News Reader

Show HN: TT911 – Text 911 your location and name without unlocking your phone

smsto911.com

31 pointsosipovas32 comments

My toddler is going to love this app. He can order up flashing blue lights and a siren any time he wants!

We built the app to enable users to text 911 dispatchers their exact location and name at the right time — without unlocking their phone.

Earlier this year we saw Text-to-911 launch in several areas around the country with the effort to improve public safety.

It works wherever text to 911 is available (predominantly Chicago, Dallas, Indiana, Maine, and Vermont).

We would love your feedback and thoughts.

I like it, but i'd be too afraid of accidental button clicks to install it. (May not be safe, but it is what it is)

When the location isn't available, and they try to use it and you state that it isn't available...

Can you at that point have an option to initiate a call to 911 (through a locked screen)? I'm not familiar with android development.

No. On Android only System Apps have permissions to place a call to 911 or other Emergency Numbers (112 etc). We can definitely dial the emergency number, 911 in North America, for the user (they still need to unlock the phone to press the "Call Button" unfortunately.

With feature phones it was possible to dial <emergency> by feel, without removing the phone from my pocket (or making it obvious what I was doing). I'd love to be able to fire off an emergency SMS without having to expose my phone or having to use the touch screen.

In my head I'd love to be able to morse-code SOS on the volume keys as a trigger. Possibly once on the v-down, phone vibrates and then again on the up to stop accidents.

I wonder if the UK can accept SMS messages on it's 999 or 111 service... love the app though :)

OK. So here is a question that is bothering me: how did you test this? How did you ensure that this application will reach the 911 SMS service?

They probably don't have to actually test it very much, the actual sms should be handled by the api which wouldn't be tested by them. Other than that they could just alert the 911 service that they are doing some testing during a certain period of time.

If Text to 911 is not available in your area you get a bounce back message from your carrier.

The FCC website has more information about TT911 here: http://www.fcc.gov/text-to-911

But that doesn't answer my question. How did you even test this?

911 is not a trivial service that can be tested blindly.

911 will accept test calls. You can call the office ahead of time and arrange a test call.

Alternatively you can dial 911, inform the operator it is not an emergency but a test call, and have them verify the phone number and location.

That very much depends on the locale for the latter. If I call 911 and either hang up after they answer or if you say that it is not an emergency but rather a mistake or test, they'll still arrive where I am.

Some places require you to get a test authorization in advance, others don't. Everyone allows testing though. Most encourage it.

I started Dash911 and more recently MultiAlert911. In every case we've test called a PSAP, as long as we clearly started the call with "this is a non emergency VOIP test call" we had no problems.

Can they get your location if you call 911 and hang up from a cellphone?

Where I live (in Canada), one time I did some test using a VOIP application, I have no idea why or how, but it caused a call to 911 (I didn't even saw the call...). They called some minutes later to see if everything was alright and when I said it was probably my test using VOIP that caused that, they understood instantly, like it wasn't unusual.

You're getting downvoted, but I can corroborate. I've called 911 to verify that my VOIP install is providing correct ANI.

If you terminate a PRI into your PBX, you can send whatever number identification you desire.

This is fun for mucking about with coworkers (why is my wife calling me from home in the middle of the day?!?) but serious business for 911 services. You get it right, and you test it by calling them. Identify clearly and immediately that this is not an emergency, but an ANI test and would they please verify that the originating number is xxxxx.

It's fully expected and might even be "required".

While you may send what ever calling ID you wish for regular calls, the local LEC treats 911 calls very differently and typically puts them through 911 screening tables that only allow the CLID owned by the account to pass. If another number is there, it can be stripped out and replaced with the billing number. Entirely up to the carrier.

Regular Calling Line ID to your cell phone may be totally different that Calling Line ID (ANI) to the PSAP 911 center.

Right, CLID and ANI are definitely not the same.

I'm sure there's some variation based on the switch manufacturer or software version at the CO, but it is definitely possible to confuse E911 services by misconfiguring your PBX. IIRC, without configuration (if your PBX sends no data) the network tags the call with 000-000-0000, which might trigger some kind of subscriber table lookup.

But it's also possible to send different ANI based on your originating extension. I had different ANI configured for my and a few others' home VOIP phones (connected by T1s to the PBX in the office, itself connected to the PSTN by a PRI into the router), so that a 911 call from our houses wouldn't summon an emergency response to the office in the Financial District.

Testing to 911 can be established by contacting them on the NON EMERGENCY NUMBER, and requesting to speak to whomever is in charge of test requests. They will tell you the policy and guidelines for that agency. This was NOT designed for general public testing of the Text service, and you could easily overload the center when an event is happening.

Don't guess, play it safe, but remember, if too many people start to test before a policy is set up, you are going to have a negative impact on the workload of the center, and they will discontinue the service. Not good for anyone.

How does this app find a "specific address" to send, as the website describes it? What guarantee is there that it doesn't send the cops to my neighbor's address?

We use a reverse geocoding service. The goal is to provide an address that is close to you (within 50 or so metres) which is better than the alternative of what responders currently get when you call and do not know your location.

Within the app itself you can move the pin (a la Uber/Lyft style to update the address) you want to send.

Here is a bit on the current status of Wireless Enhanced 911

http://en.wikipedia.org/wiki/E911#Wireless_enhanced_911

Cellular phones can use several technologies that can assist in providing and X,Y coordinate of the device. These can be extremely accurate, depending on the data available. Look on Wiki for 'SkyHook'. This information is NOT available to 911 on phone calls, nor dopes a data channel exist to even transmit the data from the device to the 911 center. This is why the App is useful, but the App, needs to talk to something on the 911 center side. Without that piece, there is no way to get the information from point A to point B.

Fletch, that's spot on. This is why the utility is gated to the areas where text to 911 is available (in upcoming update) and why 911 will send you a bounceback if the dispatchers aren't so equipped

For what it's worth, according to FCC data, 9 out of 10 wireless 911 calls don't have accurate location info, so it sounds like this service could be pretty useful: http://arstechnica.com/business/2014/07/9-out-of-10-wireless...

This is based on a report by FindMe911.org and it is specifically talking about issues in the Washington DC area based on data obtained from a FOIA (Freedom of Information Act) filing by that group. It is not to mean that this is the condition all over, while I will admit there is a problem with cellular phone location as the GPS signal is seriously disrupted inside of buildings due to the lack of Line of Site.

Yes, this is a problem, but this 90% error rate is based on statistical data from a particular area.

Have you thought of accidental calls?

Cellphones used to have similar emergency buttons. People sitting on their phones caused too many calls and these buttons have now disappeared. [1]

[1] https://en.wikipedia.org/wiki/Pocket_dialing

We certainly have though about accidental calls. That's why we implemented a double tap (with a timeout) as well as a countdown where the user is able to cancel. False alarms are definitely something we're keeping our eyes on.

Do you have a video showing this interaction?

Here we go!

Double Tap to Confirm: http://gfycat.com/UnderstatedRelievedIncatern

5-Second Timeout: http://gfycat.com/HeartfeltHeavyAkitainu

Double tap is probably not good enough. Maybe a swipe pattern? Left left right right or something of that nature?

good idea

It requires multiple actions, not a single press, and smartphones still have emergency-call-without-unlock functionality (at least, my Galaxy Note 3 does) like the emergency-text-without-unlock this system provides.

Sure, the old one-button emergency call feature is gone because of excessive accidental calls, but that's not what this is (even substituting "text" for "call".)

Reply on news.ycombinator.com