Tuesday, August 21, 2007

Whoda Thunk It? A Tropical Storm Hits Oklahoma!

Normally when it comes to weather events in Oklahoma, we're 1) so dry the ground is like a brick 2) so hot it seems like we can fry eggs on the sidewalk, or 3) dodging tornadoes.

So, it was surprising even to me to see people being rescued from flood waters by helicopters, boats, jet skis, etc. You may have seen this on the national news this weekend.

Well, folks, that's not normal for us. I looked online to see where the rain was heading and I saw a perfectly formed tropical depression swirling just north of Oklahoma City. There was even an "eye"!

What I was seeing were the remnants of tropical storm Erin, which made landfall in Texas and moved north, even gaining strength along the way.

Some towns got 10 inches or more of rain. At my house we got about 6 inches in less than 24 hours. There were flash flood warnings, but hey, we've been getting those quite a bit this year.

The thing that made this event even more unique and devastating (at least 6 people have died) was that the ground is so saturated, the water ran off the ground very quickly.

All this got me thinking about the local forecasts for the weekend. I think we had a 40% chance of rain going into the weekend. Even during the event, I saw no urgent evacuation warnings on TV. To be fair, much of the rain fell in the early morning hours of Sunday. Things that everyone agree on were that 1) this was a really big and unusual weather event, 2) it was largely unexpected and 3) we were totally unprepared.

By the way, there was also a tornado verified during this event!

What does this have to do with testing?

I've been giving this presentation on the Risks of Risk-based Testing lately. My premise is that even though risk-based testing is a good way to test, it is possible to miss risks for a large number of reasons. At this point, my list is at 12 reasons, but I think I have #13 ("We ain't never seen this before!")

There are some risks that we simply don't realize because we have never seen them in our context before. This weather event was an example of that in a non-software context.

In the context of software, perhaps a software application may fail in a way you have never seen before, in a place you've never seen failures before. A big part of dealing with those situation is to be able to identify them quickly and react accordingly. This applies big-time in dealing with security threats, which there is sometimes the surprise of a totally new type of attack.

As you plan your tests, keep in mind that "You don't know what you don't know" and keep an eye out for things you never seen before. As some say "Expect the unexpected." Oh, and a good set of hip waders can come in handy at times also.

No comments: