I recently finished Effective Software Testing – A Developer’s Guide by Maurício Aniche, and I really liked it. I have been coding for a long time and I think I have been writing pretty good tests for the features I have implemented. Even so, I found this book quite valuable. Particularly the chapters on how to systematically come up with test cases based on the specification, inputs, outputs and the structure of the implementation.
There seems to be an established truth in programming that code reviews find a lot of bugs. An example is chapter 18 (Modern Code Review) in the book Making Software. This is however not my experience – I rarely find outright bugs when reviewing changes. But I still think code reviews are useful. Below are my thoughts on the value of code reviews, and how to make the process efficient.
I really liked A Philosophy of Software Design by John Ousterhout. It is compact and short, only 170 pages, so it is a quick read, but it contains many good ideas. The focus is on how to structure systems to make them easy to understand and work with. The author is a professor of Computer Science at Stanford, but he has also spent 14 years developing commercial software.
I used to think that I didn’t need comments if I wrote self-documenting code. However, I have realized that I do write comments, and that I find them really useful. To see how many comments I write, and what kind they are, I wrote a script to analyze my git commits from the last six years. In total, seven percent of my committed lines contained a comment. This blog post has details on what constitutes good and bad comments, as well as more statistics from my script.
Microservices have gone through a complete hype cycle. From being hailed as the one true way to build systems, to a backlash where they are often considered too difficult and complex. In a way though, they are like most technology – there are trade-offs. You get some benefits and some drawbacks. Here are four features specific to microservices that I have come to appreciate.
For the past few months, I have been interviewing with several companies. In all cases, one or more coding tests were included. I have also been on the interviewing side, evaluating a coding test. Here are my thoughts on the process.
Here are more good programming quotes I have found since my last post.
“It has been said that the great scientific disciplines are examples of giants standing on the shoulders of other giants. It has also been said that the software industry is an example of midgets standing on the toes of other midgets.” Alan Cooper
“Changing random stuff until your program works is bad coding practice, but if you do it fast enough it’s Machine Learning.” via @manisha72617183
Choosing test values when writing unit tests is mostly guided by the need to cover all cases of the program logic. However, some values are better than others. Here are a few tips on how to pick values that make mistakes easy to spot and the tests easy to read. Plus a bonus tip on a quick way to double check your tests.
This fall I took the course Mathematical Modelling of Football from Uppsala University. It was taught by Professor David Sumpter, and I believe this is the first academic course of its kind. The main subjects covered are modelling and analysis of events (on the ball actions), movement and pitch control (tracking data), player evaluation, and match result simulations. There were also several guest lectures from (among others) William Spearman, lead data scientist at Liverpool FC, and Javier Fernández, head of sports analytics at FC Barcelona.
The tools used were Python (using Anaconda) with NumPy, Pandas and Matplotlib. The course was a lot of work, especially the assignments, but I really enjoyed it and learned a lot.
You have developed a new feature. The code has been reviewed, and all the tests pass. You have just deployed this new feature to production. So on to the next task, right? Wrong. Most of the time, you should check that the feature behaves as expected in production. For a simple feature it may be enough to just try it out. But many features are not easily testable. They may be just one part of a complex flow of actions. Or they deal with external data fed into the system.
In such cases, checking if the feature is working means looking at the logs. Yet another reason for checking the logs is that the feature may be working fine most of the time, but given unanticipated data, it fails. Usually when I deploy something new to production, I follow up by looking at the logs. Often I find surprising behavior or unexpected data.