Hiring great software developers is certainly not an easy task. To find out if a candidate is a good fit, companies usually focus on assessing an applicant’s knowledge, social skills, and intelligence.
THE STATE OF THE ART
Knowledge means that a person has a profound command of the domain, technology, and tools required for his/her job. If, for instance, you need a C++ developer for the development of an ECU (electronic control unit) for a car, you probably want somebody who is not only proficient in C++, but also has expert knowledge of embedded systems development, Autosar (an automotive middleware), CAN/FlexRay/MOST (some automotive communication buses), and probably a lot more technologies and tools. Determining a candidate’s level of knowledge is straightforward: you just need to ask enough technical questions.
Since today’s projects are usually done by teams, not individuals, social skills are very much sought after by every company. In general — and contrary to what is portrayed in Hollywood movies — software developers aren’t those pizza-eating and Red-Bull-drinking loners that reverently stare at their screens all day (and night). While such stereotypical coders do exist, they are the rare exception. These days, most professional software developers spend a significant share of their time communicating with peers, managers, and sometimes even customers. Developers need to successfully explain, present, and negotiate. Being able to do this in an open and respectful manner is not just highly desirable — it’s an essential survival skill. How to find out if a person has social skills? You ask questions and pay attention to how things are said and presented. Have candidates talk a little bit about themselves, their past projects, let them explain a difficult topic.
Many traditional companies would stop here. However, most successful, especially software-centric companies have recognized that while knowledge is important, it’s not sufficient. They want smart developers; that is, developers that come up with creative, efficient solutions and require little time to implement them. Thus, they also check a candidate’s intelligence by torturing them with riddles and logic puzzles, including those who look like they’re impossible to solve (“How Would You Move Mount Fuji?“). Others ask candidates to do online programming tests that challenge the candidate to find clever solutions to tough programming problems. The reasoning is this: intelligence ensures that knowledge is put to good use.
By now you know whether your candidate is a great developer, don’t you? If a candidate passes all the aforementioned tests, (s)he certainly has a lot of ability. But does this also imply that such a person will get the job done? Sadly, it’s a common fallacy to confuse ability with willingness.
More than fifteen years ago, I worked on project where the development team received insanely formatted error log files from the testing department. Information that would have helped us debug the problem was buried deeply in ASCII noise. Every time we received a bug ticket, we had to spend quite some time on figuring out what exactly went wrong. One day, a coworker gleefully came into my office and told me that he’d written a filter program that extracted and reformatted all the relevant data from the error log files. What a neat idea and what a time-saver! Full of pride, he showed me his code. What I was faced with was many pages of C code, written over the course of three days, full of low-level ‘getc’-based text parsing — definitely not easy on the eyes! Finally, I gave him a smile and thanked him for this great little tool. Half an hour later I went to his office and showed him a 30-line Perl listing that was functionally equivalent to his C code.
He didn’t know about regular expressions and neither Perl nor any other scripting language. All he knew was C. He clearly lacked knowledge. He was a lousy software developer, right?
As a matter of fact, I think he was a very fine developer. None of the other team members (including myself) — despite knowing about regular expressions and Perl — could be bothered implementing such a productivity boosting tool. Instead of lamenting, he just did it, even though it must have been painful for him. He didn’t waste three days because of his incompetence. His persistence saved hundredths of developer-days down the road. He did it because it was the right thing to do.
Of course, he was somewhat embarrassed at first. I told him that I reimplemented his program to demonstrate the power of regular expressions and dynamic languages, not to insult or humiliate him. Guess what? A couple of days later, he told me that he did a Perl tutorial in his spare time, because he was sure it would save him lots of time and effort in the future. He never defended himself and he didn’t make lame excuses — he acted like a pro.
This is a guy you certainly want to hire! He possesses two traits that are much more valuable than knowledge: persistence and professionalism.
PROSPECTING FOR GOLD
To me, persistence and professionalism are not fundamental personality traits; I strongly believe that they’re the by-products of a root trait that I call “professional software passion” (PSP). Developers who possess a high degree of PSP are self-motivated and in general enthusiastic about their craft, they commit to life-long learning, are able to take a lot of hardship and always work towards fulfilling the business goals of their employers and clients — not just the short-term goals — especially the long-term goals.
Unfortunately, typical hiring processes completely ignore this critical trait. A certain level of intelligence and social skills is crucial, no doubt about it. Knowledge, however, can and will be obtained as long as PSP exists in a developer. PSP ensures that the job gets done. I would pick PSP over knowledge any day.
I will explore the topic of “professional software passion” as well as strategies on how to detect it in candidates in future posts. Stay tuned.