Maybe I was quite naive about peer reviews but my experience previously was that it was a natural process to have QA/testers involved in code reviews alongside developers. Whenever a new feature or a bug fix was implemented, before the code was checked in, the developer would set up a walk through with another developer and a member of the QA team. The team I was part of was quite a mature engineering team with a defined coding standard which was ingrained into the team. Never was a second thought given to the fact that development would commit code without showing it and talking through it with QA. This would provoke discussions around how to test the features and whether the development written unit tests had enough code coverage to QA’s satisfaction. Moving to a different team, I have seen that this isn’t necessarily part of the process, peer reviews happen between devlopers, it has never been considered to involve QA.
I have read a lot about this subject and actually, the level of maturity around code reviews of this first team is relatively rare in the industry. So, why is it so rare? I guess it depends on perspective and the level of testing being done:
- If ‘black box’ testing is the main form of testing, then I guess not knowing about the code is the ‘right’ thing to do?
- If ‘white box’ testing is used, then knowing the functionality you are testing is paramount, other than functional specifications, the best way of seeing how the area being tested works is to review the code.
All testing I have ever done has been a mixture of both of these, there has obviously been a level of testing which I can pull straight out the user guide/functional specification and others where I have needed to know intricate details of the code to be able to ensure I am testing all feasible code paths.
I have always found it beneficial to be involved in peer reviews, even if I don’t say anything during the review, but just soak in how the code works and write down ideas for testing. But usually, I will ask questions such as “what if i entered this here?”, “what if i did this?”, using the meeting to also force the developer to think about their implementation, rather than just plodding through code line-by-line.
So why is it so alien to some teams to involve QA? Here are some of my thoughts on some common phrases used:
- QA don’t have the skills to review code – Not every QA resource will know the syntax of the particular language, but does that mean by sitting with developers and understanding how the code works, they won’t find issues or raise questions which provoke the developer to improve their code?
- Having QA involved will delay build/release – Only if you treat QA as a separate entity to development and do separate peer reviews. If they are involved in same peer review, then it shouldn’t make much different. We are here to prove the quality of the work, not be a hindrance.
- It’s only a one line change, why would QA need to review that? – On one hand, yes it is only one line but the context of that one line could have an impact on some existing testing, or just having QA aware of the change could be useful.
I’m not for a second suggesting that development teams may be dead against the idea, but I think as we move to an increasingly ‘agile’ world, separating Development and QA out at this level needs to change, we should be promoting a ‘One Team’ approach where value is provided by everyone involved. QA can bring value to code reviews. Quality should be built in at the start and the earlier this can be proved, the better it is going forward. It needs to be clear that performing a critique of the code, is not performing a critique of the developer.
Some quick wins may be needed to win the development team over:
- Read up on the functionality before you attend the review, so you have a basic understanding of how it was described to be developed
- Ask logical questions
- Discuss options for testing
- Flip it and have development review tests (share both activities)
Baby steps are needed for progress, find a way to get involved for some small tasks to start with and build up trust with the developers that you are not just there to be a pain in the backside but actually, if you work side by side with the development team, it will improve the product delivered.
Any thoughts on this topic would be most appreciated.
4 thoughts on “Overcoming the Resistance – QA Involvement in Peer Reviews”
Excellently put in my opinion. I’ve worked 7 years in test roles and not knowing the code is not an issue. We are all interpreting the same requirements and specification docs which are not always the best! Even in my first couple of months I found that attending Dev team meetings allowed units for testing to be scheduled in a more sensible order. There will always be something useful picked up. IT solutions can be quite complex and we all have useful contributions which should be seen as positive and in no way undermining. IT is best when we are all in a collaborative mindset to reach the common goal.
I supported many software product reviews representing test/QA. I was always well received, but then I added value that the whole team came to respect by providing input from a different system-test-QA view point. I used checklists so I made sure the reviews were consistent and everyone came to know my checklist to the point where they made sure they were always addressed. Everyone won.
Very interesting! I had never thought about this but it makes so much sense.