Well, I'm trying to get home from a class (Adding Value to QA and Testing Processes) I presented in the Boston area this week. It was a great session and I really enjoy teaching people who are engaged in the topic. I was so glad to be able to actually get to the class because in Oklahoma City we were hammered last Sunday by a devastating ice storm.
There were lots of trees destroyed and 700,000+ people without power at some points. (My thanks to the folks at OG&E and the people who came in from outof state to help us! God bless you!) In fact, my flight was one of the first to take off Tuesday afternoon, as the airport had been pretty much shut down most of the weekend.
As it turned out, I was jumping out of the refrigerator into the deep freeze!
As those of you in the U.S. know by now, Boston was hammered by a snow storm on Thursday (Gee, it seems like last week already!). Here's what made it interesting.
Despite the warnings and planning, most of the state was paralyzed. In fact, the warnings may have actually contributed to the problem in a major way.
It took me 4 and one-half hours to travel about 16 miles on the highway. The highest speed I reached was 15 mph, and the average was about 4 mph. I was not alone. In fact, I fared pretty well as compared to some.
The crazy thing was that the snow wasn't that bad. Listening to the radio was a riot. People who have lived there all of their lives were ranting about "what's happened to us here?" I'll admit, as a "Okie", I was thinking that even we handle winter weather better that this. (No offense intended.)
Here's what I think happened and it applies in other situations as well. Warnings went out the day before that the snow would hit during the evening rush hour, so people should plan on leaving work early. So, as people are (including myself), we waited until it started snowing around 1:00 p.m., then left. It was snowing about 1" per hour, so it didn'y take long for it to start getting messy.
However, because EVERYONE left work early, rush hour started at 1:00 instead of 3:00 or 4:00. So, the snowplows couldn't move the snow because of all the people trying to get through the snow. What irony! You might say, the perfect storm.
As much as the weather caused a problem, so did people's response to the weather. I think we may have been better off just giving the forecast and letting people figure out for themselves when to leave work.
We see the same thing happen with software and systems. Consider the web site to buy concert tickets that only gives people two minutes to complete the transaction. Because of the heavy load seen when the tickets go on sale, it's hard to get the response time to finish in two minutes! (As you can tell...I have experience in this scenario.)
When thinking about system performance, give very careful consideration about what people may do that could actually aggrevate problems (like hitting the Enter key multiple times just to make the system go faster). That's why good system performance (and weather response) needs to be well thought-out.
While we can't control what people do, we can influence what they do.
OK, lesson learned. If all goes well, I should get home to Oklahoma City just in time for another snow event. I think I feel a "sick day" coming on.