You are viewing a read-only archive of the Blogs.Harvard network. Learn more.

Teaching Class on Git

I spent some time this past week teaching a class on Git for the Treasury Department. We had a lot of fun. I used Digital Ocean to spin up an instance for the students instead of the AWS instances I usually use. It seemed to work well and in some ways it was simpler than using AWS.

Mark Shead Teaching Git Class
Mark Shead Teaching Class on Git Version Control

Can You Measure It?

For years there have been two elementary schools in Fort Scott. Eugene Ware Elementary School on one side of town and Winfield Scott Elementary School on the other. This year the school board voted to make a change and all of the same grades at the same school. This means regardless of where you live, you have to go to the same school for your grade. This was presented as a way to save money.

I can see some ways that it might save on the number of teachers. If your maximum class size is 20 and you have 30 students in a particular class at each school, then it will require four teachers with two at each school. Each would have an average of 15 students. If all of these students are at the same school, you could have classes with 20 students and only need 3 teachers. So the combination may be able to reduce the number of teachers required–but only at the expense of larger class sizes.

While I applaud the idea of trying something new, I was very disappointed that there was no talk about it being done as an experiment. These type of changes have a way of being much more complicated than they look on the surface and it is impossible to tell if it is really going to save money until you actually try it.

This isn’t just a problem with school systems. All kinds of organizations make big changes with no clear plan for how they are going to measure the results. Continual improvement is the process of changing and then measuring. You keep the changes that are beneficial and jettison the ones that are not. This only works if the measurement aspect is “built-in” to the change process. With good measurements and a true commitment to roll back anything that doesn’t work, you will make progress even if you just make random changes.

Obviously, well planned changes with some intelligence behind them will move you forward much faster than just making random decisions. However, it is a mistake to assume that intelligent decisions can substitute for identifying and rolling back the things that don’t work. The power is in getting rid of the things that don’t perform as expected.