I've got a few quibbles with the article: (1) It's not obvious to me how this, or any app could have prevented the rape the story describes. And (2), the 'app' isn't an app at all, but a website and SMS based service. So yeah, don't bother searching Google Play or iTunes, you won't find anything there related to site in question.
Still, the article gets points for mentioning the service. I like Kitestring as both utility as well as a case study for entrepreneurs.
The service is quite simple: you report that you're starting a trip of a certain duration. At the end of the trip, the system checks in with you. If you send the system your check-in word, you're all clear. If you don't respond, or send the system your duress code, your emergency contacts are notified. The entire webapp consists of one page where you can set your various 'words', maintain your emergency contact list and customize the message to go out.
The system would work well for those walking home late at night, heading out on a trail for a trail, or even when stopping for gas in a shady neighborhood. I could see using it before heading out on an epic hike or monster bicycling trip. It would provide an automatic backup in case I got lost or delayed. I could even imagine parents using this as a sort of pop-quiz for their kids: they setup a trip on their behalf, and if they don't check in, they get a phone call and tracked down.
The system doesn't do anything fancy with the user's GPS or have extensive options. But I think that's a good thing. The simplicity means that it works, and you'll use it.
As an idea guy / programmer, I'm always going on and on with folks about how they can turn their big-huge-awesome idea into something that they can start building today. I have no idea who's behind Kitestring, and I have no idea what their philosophy actually is, but on the surface it completely matches up to what I tell my customers.
I can see the pitch: let's revolutionize personal security! Let's make it app based, location based and fault tolerant. Let's use a heart rate sensors to detect if the person really is in duress. Let's use the camera to snap a picture of the assailant. Let's integrate this in with 911 so the police can be seamlessly dispatched. Let's develop an advanced algorithm to detect false alarms from true crises. And I say terrific! Let's do it all! But let's start with version 1.0.
Version 1.0 needs to be small (it'll cost you less, be faster to build and more importantly, lower risk) but mighty. It needs to capture the very essence of the idea without having any extraneous features. Ideally, it could be used in a number of contexts, allowing people to re-purpose the system in ways the creators never thought possible. It needs to deliver true value, and pique users interest so that they'll give you feedback on what to develop next. And it needs to be something you can start on today without anyone's permission.
I believe Kitestring has nailed these essentials. Is it perfect? Of course not. But they appeared to have hit all the above points. If you're looking for inspiration for building out your idea, they'd be a good place to start.