Software quality at 10,000 feet

I’m returning home from India. I talked, I taught, and I discussed software testing with a lot of the testers at the India Development Center (IDC) in Hyderabad. The trip had its ups (great people and great sights) and downs (a bit of food poisoning on the last day), but overall was a fantastic experience.

About 45 minutes into the flight, it was announced that the entertainment system was “acting funny”, so they were going to reset it. This, unfortunately meant that all of the pre-entertainment safety and informational videos needed to restart. Unfortunately, the reset didn’t work, so the flight attendant announced that they were going to “do something called a hard reset”. An interesting issue with system integration causes all seat power(electronic controls on the seat as well as reading lights) to be off-line until the system reboots (about 10 minutes). I watched the system reboot (appears to be a linux system), and it looks like it’s doing better.

The in-flight entertainment is an on-demand system. I like flights with this feature, because I hate trying to time the beginning of a move after sleeping or eating. However, the flight attendant made another interesting comment about the system; she said that the first time the system was used it would take 5-10 minutes to respond. Furthermore, if the controls were pressed repeatedly in an attempt to “speed it up” (or to see if was responding), that the server would reset, and there would be a fifteen minute delay before the system could be used from that seat again.

I could hypothesize on the customer focused design techniques (or lack thereof) that whent into developing this system, but I bet that for the crew, it’s apparently “good enough”. Providing entertainment isn’t a key feature of the air transportation industry, so they are willing to deal with this sort of functionality. I’m not much of a movie watcher (especially on seven-inch screens), so it really doesn’t matter to me. I’m a bit concerned on the dependencies of the different systems – generally, less dependencies lead to higher quality, but I can only assume that business needs (cost) are at play here.

It turns out, that after several reboot attempts, the flight crew has given up on the system, and this will be an entertainment-less 9.5 hour flight. Luckily, I have some work (and sleep) to do, and there’s a trickle of laptop power at my seat. There are two things I’m bummed about. The first is that I dropped my mp3 player on this trip and it bricked on me (old school HD based system), so I will be stuck listening to airplane noise (I guess I finally have an excuse to buy a zune). The second is that I don’t have a camera – I’m in the first row, and I’m looking at the “control center” for the entertainment system. At first, it only looked familar, but I quickly recognized it – it’s running Windows 3.1. There’s even an old school GPF on the screen right now.

Comments (2)

  1. There are a number of rather interesting problems mentioned in the story. "An interesting issue