Trapster: A Nice Idea Poorly Implemented

There are lots of uses for an iPhone and it’s good to have one in your car, especially in the case of Trapster, an iPhone app that aims to alert you about speed cameras and other gear out there in the streets that can get you fined. It’s a brilliant idea, but let’s see if this app does what’s promised and whether it performs that task well enough.

How does it work?

First of all, even though the app’s description doesn’t clearly mention it, the app alerts drivers based on a user-populated database; the iPhone does not have the necessary equipment to detect speed cams! What that means is that you get alerts only for those traps that other people have reported, and that data may be inaccurate or even completely false. There is though a means of rating each trap and the app calculates a confidence level for each one, but still this is all based on people’s good will.

One case in which the user-centric database can be of no use is that of systematically submitted false alarms. Some anti-trap-alerting group, for example, could massively submit numerous traps that are not real, thus making the app useless for a driver that would constantly get alarms for cameras that are not there. But assuming that there are no such groups (yet) and the authorities would never come up with this, false alarms shouldn’t be a big issue, although some of them surely exist.

Jeez, where’s that camera?

The main problem with Trapster is not the data, but the way it reports traps to the user, who in this case is, the car’s driver. This is where Trapster fails to deliver in a practical way, for many reasons. First of all, the reports are using a synthesized voice that not only sounds rather ridiculous, but is hard to notice at times over the noise in the car. Also, don’t hope that you can scroll the map as in the Maps app -no, no! The map stays fixed and you have to hit a “center” button to re-position your car in the map.

But most importantly, the problem is that the algorithm for trap detection is based purely on proximity and doesn’t account for the car’s direction, thus alerting drivers for traps that are monitoring the cars that are going to the opposite direction! This can be rather frustrating if you’re driving in quarters that you’re not familiar with (but that’s the point of this app, if you already know by heart where the traps are, it’s of no use, right?) since you’ll be alerted for traps that aren’t meant for you, but you’ll still be slowing down and wondering where in Earth that trap is supposed to be.

Another problematic area that can be improved is when a user submits a trap, since the map problem mentioned above doesn’t make this too easy for a driving person. I think it would be a good idea to let users log in to the site when they get to their computers and fine-tune the data they entered in a hurry while driving, or even add some more details about the trap.

Bottom line

Trapster is a nice idea that is poorly implemented, at the moment. However, it can still be quite handy and save your from a speed ticket or two, so it’s good to have the app running in your iPhone while driving in a highway. Knowing it’s problems and limitations will reduce your frustration, and keep in mind that this app uses data generated by people like YOU - so don’t forget to submit the traps you see, or rate the existing ones!

Wishlist

Here’s a list of a few ideas that would make this app much better:

• Detection algorithm should account for the vehicle’s course (e.g. create a vector simulating the vehicle’s track and associate that with a user-submitted trap, then check against this before alerting users).
• Use a clear and loud human-recorded voice for reporting traps.
• Calculate the vehicle’s speed and use audio cues to report how close the vehicle is to the trap (i.e. a driver’s that’s moving too fast should be notified earlier than one who’s barely above the speed limit).
• When such audio cues cease, drivers would also be informed that they have passed by the trap and are no longer in it’s range.
• Fix the interface, please (draggable map, zooming with gestures, auto-following the vehicle - you know, like the Maps app!).
• Allow users to refine trap data through the website.