Last week I joined 30 Days Of Testing -challenge that was organized by Ministry of Testing. And I have to admit it has been fun. Besides that I have found myself doing things I should’ve been doing all the time, I’ve also done something new, too.
First of all the ‘Listen to a testing podcast‘ was very revealing as well as fruitful. I’ve been now listening to the Test Talks while riding bike to and from work and found out new ideas, new ways to think about testing. It is really refreshing to notice that, even though I’ve been working as a single tester (and I’ve done that on purpose, also) quite a lot for the past years, there’s a community out there. People thinking, if not like me, but at least similar kind of issues and situations that I am facing in my profession. So to say, it has been a blast to notice that 😀
Yesterday’s challenge was to do a ‘Crazy Test’. So, what I did was the following:
- Connect iPhone to the car with Apple CarPlay
- Use Siri to start music (‘Play music from artist Metallica‘)
- Wait until Siri gives you a ping that it has received the command (but before it confirms what it is about to do)
- Put on reverse gear
Now, my expected result would be that I could reverse in the car by watching the rear camera and listening to music (at this example, Metallica). Of course it didn’t do anything like that. I received the ping from Siri to indicate that it had received my message, but when the rear camera was switched on, the command was weirdly gone. I could reproduce the behavior even when making a phone call request.
Now one might thing it is a safety feature. But I can’t think it that way, for me it feels like a bug. Software is not working the way I expect it to do. If I turn on the phone and call manually and start to reverse, the call is connected. As well when it comes to music. I can pretty well listen to music (even Metallica) and reverse at the same time.
Now, is this a race condition, or something else? Where is the problem? In Siri/Apple CarPlay? In the API between car and carplay? Has this scenario been tested? Does not seem likely. Or if it has been tested, perhaps the results have been ignored in order to get the software out in time. Or the responsibility of the behavior has not been clear; should it be car that takes care of this, or the phone?
You might also say (and I wouldn’t argue against that) that this wasn’t particularly crazy test. More or less exploratory boundary/edge case. I was not driving my car sideways on the ice, or 110 Km/h on the highway. I don’t actually care. It was crazy enough for me. Plus that I noticed something new, learned something new, which I suppose should be the main point in life and work altogether. i mean, you can never be crazy enough, not unless you learn more and get crazier.
And I’d like to automate that test. For real 😀