Recently I wrote a Python test which I was going to use during interviews for hiring developers. I never got to use it sadly as I decided to leave the job it was for, let’s just say the company had a few issues…
The test asks a developer to complete a few tasks based around list and dictionary manipulation. I shared it on Twitter and got some interesting feedback. A number of developers / friends thought it was a little too easy.
it's almost too easy. I don't think you can figure out if someone is mid/senior from this— @r4vi (@r4vi) October 11, 2018
This was fair feedback but missed the point of the test slightly as it was never meant to be difficult. The reason is a developer test or assessment, I prefer the term assessment, is not checking to see ‘if’ someone can write code and solve a problem. Instead it assesses ‘how’ a developer writes code and solves a problem. And ‘how’ they solve the problem defines where they sit on the Junior to Senior spectrum.
Personally as a manager I measure developers as follows:
- Junior: Can solve a problem.
- Mid: Solves a problem well, shows knowledge of principles, adds some comments and tests.
- Senior: Solves a problem well, shows knowledge of principles, adds full comments, writes tests, complies with a code standard and applies static analysis.
My standards are of course arbitrary and not everyone will agree with them, however they can be validated by a relatively simple assessment. And in reality the main differences between junior and senior developers are quality, professionalism and experience.
A good assessment is one which is open and gives a developer enough rope with which to hang themselves. If they don’t add comments this tells you something important about them. So ultimately the assessment does not need to be complicated or difficult to complete.
Also assessments are carried out in a developer’s free time. So it’s unfair to ask a developer to waste hours on something that may not result in anything positive for them. So an assessment should be short, it definitely shouldn’t take more than an hour to complete.
A great example of a good assessment is FizzBuzz, it is a relatively simple assessment and there are numerous ways to solve it.
Write a program that prints the numbers from 1 to 100. But for multiples of three print “Fizz” instead of the number and for the multiples of five print “Buzz”. For numbers which are multiples of both three and five print “FizzBuzz”.
To provide an example of how you should approach development assessments I have created a Python repository which solves FizzBuzz. I’m not going to claim my code is perfect, there is certainly room for improvement. You will though note a few things:
- The code has tests and I’ve achieved code coverage of 78%.
- The code complies almost perfectly with Pylint which imposes the PEP 8 standard.
- The code is run against the Radon static analyser and achieved an A grade.
- The code has comments and documentation.
- The code is run against a Continuous Integration pipeline.
By completing the above steps I can prove with some level of certainty my code works. It also shows a professional approach to development with a focus on quality. And it proves I have a reasonable level of experience.
If you are a developer looking for work, like myself currently, this is how I’d advise you approach any assessment you are given. And if you are an assessor / employer I’d remind you your assessments don’t need to be too complicated to tell you a lot about a developer.
And there certainly isn’t any need to whiteboard anyone… ;)