Toll offices, which handled long distance calls, were computerized by then. They could be upgraded to deal with 911 routing. This requires a database to map phone numbers to the appropriate Public Safety Answering Point. The idea of having a computer with a disk drive in the call chain was radical at the time, but it was made to work. The original computers for this were Western Electric 3B machines, set up as duplex pairs.
The toll office 911 system became a problem with the breakup of AT&T. AT&T retained most of the toll switches, while the local exchange carriers got the local offices. (Now, of course, with weak antitrust enforcement, they've all merged back together again.)
The article is correct, advancing technologies are not reflected by additional capabilities by PSAPs, DISA's or ILEC's (The ppl that foward your call to a responder). It was only recently that IP based 911 sending of ANI/ALI (Telephone/Location data) data has been implemented on a large scale. In Canada, it ran on an old packet switched network for ages!
Their are different flavors of 911. E911 (Enhanced), V911 for VoIP phones, and recently the addition of Wifi based calling among others. The original 911 systems was designed when phones were static, they didn't move. Nowadays, with cell, voip and now, wifi devices, theirs no telling where the call (device) originates from. Yes, most send long/lat data, but that is based on triangulation of cell towers and not accurate enough. To further complicate things, long/lat doesn't take into account altitude. In a urban setting, the responders maybe at ground level wondering just which building and floor originated the call (yes, their systems to deal with this but not widely implemented, and, failing that, if the caller is unable to speak, you have a larger issue).
Most responders don't even have the ability to map a long/lat. So at this point, the accuracy is moot.
All that being said. The E911 system works well. Cell/VoIP/Wifi, not so much. The call will generate a response, but it's hardly efficient.
NENA folks also wanted the FCC to mandate an IP location system to be available on every Internet connection - what a privacy mess that'd be, when any app with IP access could get your exact location down to the apt number.
I trie experimenting with what I named "Advanced 911". For calls to places where we couldn't deliver location info, we'd prompt the answer to press a key to have us speak the info. But getting adoption for that was super difficult.
Also, the stress involved in those jobs, wow. I thought I could handle stuff, but auditing problem calls where people were dying, panicking, and didn't know where they were - that's traumatic. Also made me really dislike how cavalier 911 is handled by the telecos.
I'll vouch for the stress of the job. If I f'up. Somebody can die. Same thing for the operators. They get 3 months of training before handling calls. Most wash out before their training is completed. Most, shortly after that. The churn is unbelievable. We hire constantly. All operators max out their sick/vacation days. It's an unpleasant, boring, terrifying job.
This is genuinely surprising to me. The operators having trouble makes sense, because phones are something of a rolling disaster, but responders aren't using GPSes that can do this?
Imagine a situation where your traveling down an interstate out in the boonies. Theirs a wreck. You call 911 from your cell. The responder will receive an approximation of your location, likely the closet cell tower. From that, and before being able to send a responder, they must determine WHO to call based on your location. Just because your closer to town B than town A doesn't mean town B gets the call. Town B may forward your call to town A. Then, you have to determine where to dispatch too. It's doubtful you know the closest mile marker. Only your direction of travel, and the last area that you can remember.
It's strictly a best effort.
999 in the UK (from 1937) was chosen because public payphone could be easily modified to make it a free call.
111 in New Zealand (from 1958) was chosen because the system was implemented using British Post Office equipment that already supported 999. But New Zealand phones pulsed in reverse so 111 on a New Zealand phone produced the same pulse as 999 on a British phone.
000 in Australia (from 1961) was chosen because 0 was already used for trunk access. On an automated rural exchange, 0 would connect you to a main centre. In remote communities it was 00. This meant that dialing 000 through an existing remote exchange would at least connect you to an operator in a main centre.
The freight elevator shaft had a shattered wooden freight elevator carriage at the bottom of it. We took the stairs.
Somebody would have to be very motivated to push through turning an interconnection point like that off. The politics involved in doing so would be extraordinary - even if there was common agreement from the start that it was time to let it go. I don't even know that you really could initiate a project like that from within a Telco.
It is a bit surprising that it was common knowledge that part of 911 ran out of there. While it's not "secret" information, the purpose and location of most similar properties are typically not openly advertised. Most people don't even know where their local CO actually is.
They basically gave up. Though the telecoms are a duopoly. Only working 911 system is in Davao.
Everywhere else you have independent short codes like 117 which may not be answered. Ironic for most of the BPO is outsourced call center agents.
Then the second issue are jokes,
http://news.abs-cbn.com/news/08/01/16/dial-8888-911-govt-ope...
"Hotline 911 received 2,475 recorded calls, between 12 a.m. to 7 a.m. Monday. Out of the total, only 75 calls were legitimate while the 1,119 were dropped and 1,356 were prank calls."
In other news, I've not yet needed it.
Many projects have their own snowflake build systems, web APIs have custom authentication schemes, multiple federated login systems...
Don't get me wrong we see a lot of consolidation, the world is getting better :)
I was in Japan earlier this year, and happened to be watching TV when the earthquake warning system cut in. It gave about 5-10 seconds of advance warning before an earthquake hit. Enough time to move out of the way of anything heavy. It's quite impressive.
If anyone's curious, this is what it looks like: https://www.youtube.com/watch?v=lAxxZkpV0HI
(The earthquake early warning cuts in at 0:15, and a tsunami warning automatically cuts in at 2:34... though most of the tsunami warning is cut off at the end. This is what the full earthquake+tsunami warning sequence looks like: https://www.youtube.com/watch?v=qMNu8Y1jIlU)
Yep. I was a 911 dispatcher for a while and can confirm this. I worked for Brunswick County (NC) 911, which borders New Hanover County (NC), Columbus County (NC) and Horry County (SC). Cell phone calls were usually sent to the right place, but if you were way out in the boonies on highway 211 near the Brunswick / Columbus line, if was about 50/50 which 911 center would get the call. And we'd occasionally get a call that was dialed in New Hanover County, Horry County, and - rarely - even further away. So on those calls, we had to try and quickly determine exactly where the caller was, so we could relay the information to the other county for dispatch. Sometimes if we weren't exactly sure and the call seemed to be near the line, we'd dispatch units and request the neighboring county to dispatch theirs as well, and then let the responders figure it out when the arrived.
I took a call once that, as far as I could tell, turned out to be several counties over... something like Pender or Duplin County. We didn't even have their contact info logged anywhere, since it was thought (at the time) that there would never be any need for us to contact, say, Duplin County. I had to call New Hanover County, ask them if they had the number for Pender (who they border), then call them, etc. The term "clusterfuck" comes to mind, but luckily that only ever happened once that I can remember.
Note that this was in the mid 90's and I'm not sure if things are better, worse, or the same now.
Are you saying the GPS they mandated in cell phones ten years ago, ostensibly just for this purpose, doesn't actually do its job most of the time?
I've actually had this happen in an emergency. I can't tell you how dismayed I was that my phone crashed when I called 911.
After that experience I switched to iPhone. However, since then, I've experienced the iPhone iCloud Phone.app crash, which also denied me 911 service (although fortunately I noticed the problem and resolved it before needing 911.)
The basic issue is that CSS colors are defined in terms of relative brightness of the monitor, and not within of a proper colorspace. Most browsers assume sRGB, but many users don’t even have that.
In sRGB, a contrast of (102,102,102) to (255,255,255) is a very pleasant, and perfectly readable contrast.
But many users are on shitty 6-bit color depth 100 nits or less displays.
Changing contrast in hardware down is always a shitty option.
The solution is to handle this in hardware — artifically increase contrast and destroy image quality for people on shitty systems.
On the other side of the balance, users perceived "9" as taking aaaaages, partly because they're bloody users and never satisfied, and partly because in emergencies perception of time gets tweaked. Anyway, 999 it was, and remains.
> The 9-9-9 format was chosen based on the 'button A' and 'button B' design of pre-payment coin-operated public payphones in wide use (first introduced in 1925) which could be easily modified to allow free use of the 9 digit on the rotary dial in addition to the 0 digit (then used to call the operator), without allowing free use of numbers involving other digits; other combinations of free call 9 and 0 were later used for more purposes, including multiples of 9 (to access exchanges before STD came into use) as a fail-safe for attempted emergency calls, e.g. 9 or 99, reaching at least an operator.
Or we simply fix browsers.
Because if it’s a choice between blinding users on most mobile devices, and most higher end desktop computers, or blinding those who have cheap hardware, most developers will choose by who is actually willing to pay.