Skip to main content

The arrogance of regression testing

Lets assume we know that our software is not perfect. How can it be? Its complex, mortals created it and we don’t have enough time to test every execution path & environment – so we could never be sure anyway. This is Ok - this is normal, testers deal with this situation every day.

This tends to be a typical scenario... Our team has been working on some new features. They’re looking good, initial teething issues have been fixed and the new features are considered worthwhile enough and bug-sparse enough to be released into the wild.

This is where things can get a little awkward. The team member’s opinions are often split across a wide spectrum. The relatively minor perceived impact of the work leads some to conclude that the work is ready for release as is.

Other team members, who are possibly twice shy from previous ‘minor change’ induced problems, argue for a comprehensive ‘regression test’ of the software. There is usually a range of views in between suggesting for example only ‘regression testing’ directly affected systems or those associated with the changes etc.

The oft-stated concern is: we may have broken something. Our new code might be good, the existing code might be even better, but what about emergent issues caused by the new ‘system’ we’ve created?

A common compromise proposed by teams and customers is generally translated as ‘test that it [the core features etc] still works’. This may sound reasonable, intelligent and practical. But it just doesn’t sit well with me. My unease stems from more than the assumption that testers can prove it works…again. The concept of regression testing seems arrogant and it even, worse it seems wasteful.

Arrogant? Regression testing assumes that we tested the application so well before that we found all the important areas of ambiguity and bugs. It’s like saying that when Microsoft released a patch to Vista that they only needed to ensure it was up to the high standard of the original Vista release.

The root of the problem here is a bias. We start out with the perception ‘Our system is great’, and then lets check (sic) it still is. The Congruence bias is a powerful motivator to not upset the perceived status quo. We stop looking for problems that we don’t think are caused by the new changes. Whereas we might investigate these ‘issues’ in other circumstances, we don’t even think to look deeper unless they are practically labeled “BROKEN BY THE NEW CODE”. How many issues are overlooked?

Wasteful? By misdirecting ourselves from the system as a whole – we are missing an opportunity to test again. A second chance to find those bugs we didn’t find last time. We could be capitalizing on another chance to learn new and old areas of the application.

When I’ve been in this situation I’ve also fallen foul of the biases at play, I’m human. But as a tester, I’ve had to come up with a few strategies to help remedy the problem.
  • Firstly, just be aware of the bias - don't let it lead you blindly.
  • If you’re lucky enough to have another tester in your team, split the problem with them. One of you can test the system as a whole, the other the ‘affected areas’.
  • Deliberate opposition. A technique I use when I’ve got a definite checklist of the affected areas. Deliberately pick things not on the list, or that are the direct opposite of what is on the list. E.g.: If the change affects Logging in as user X, What does logging out as user Y do? Or can you avoid being logged in all together?
  • Randomness. Choose a path or some data at random. For good randomness is a useful source.

I find the above a useful means of breaking out of the tunnel vision of regression testing. They give you new paths to follow, that if explored can often yield new and old bugs.


  1. Good thoughts!

    I was recently asked how I could "guarantee" that the system still works after a sw change (by use of an automated regression suite). To which I replied that I couldn't guarantee anything unless the guarantee stated that the narrow "paths" traversed by any automated suite would be the exact same that every possible customer would use (same starting states, exact same behaviour wrt timing between actions, data transmitted (being exactly the same), user provisioned data in the databases, etc, etc.).

    A very confused and disappointed face greeted me.

    The upshot was that I asked the person to be careful with language - try and be precise even - and definitely don't promise anything that you don't understand the implications of... I could tell the persons lots about the sw and the testing of it, but I'd probably never use the word "guarantee".

    Talking about testing, what it means and its implications is a big effort! And it's something that testers need to devote just as much time, effort and understanding to as any other part of their repertoire.

  2. Thanks Simon, Yes I've been in your situation also, It can be a challenge to manage expectations. Using the 'right' language can help a lot. I also find keeping a mental list of 'similar issues' can be useful. Real failures often speak louder than words.

    For example: "You're right this looks like a simple change, But remember when we released 'The XYZ', and that was a change to the same sub-system. It turned out to affect 'The MNO' - badly"


Post a Comment

Popular posts from this blog

The gamification of Software Testing

A while back, I sat in on a planning meeting. Many planning meetings slide awkwardly into a sort of ad-hoc technical analysis discussion, and this was no exception. With a little prompting, the team started to draw up what they wanted to build on a whiteboard.

The picture spoke its thousand words, and I could feel that the team now understood what needed to be done. The right questions were being asked, and initial development guesstimates were approaching common sense levels.

The discussion came around to testing, skipping over how they might test the feature, the team focused immediately on how long testing would take.

When probed as to how the testing would be performed? How we might find out what the team did wrong? Confused faces stared back at me. During our ensuing chat, I realised that they had been using BDD scenarios [only] as a metric of what testing needs to be done and when they are ready to ship. (Now I knew why I was hired to help)

There is nothing wrong with checking t…

Manumation, the worst best practice.

There is a pattern I see with many clients, often enough that I sought out a word to describe it: Manumation, A sort of well-meaning automation that usually requires frequent, extensive and expensive intervention to keep it 'working'.

You have probably seen it, the build server that needs a prod and a restart 'when things get a bit busy'. Or a deployment tool that, 'gets confused' and a 'test suite' that just needs another run or three.

The cause can be any number of the usual suspects - a corporate standard tool warped 5 ways to make it fit what your team needs. A one-off script 'that manager' decided was an investment and needed to be re-used... A well-intended attempt to 'automate all the things' that achieved the opposite.

They result in a manually intensive - automated process, where your team is like a character in the movie Metropolis, fighting with levers all day, just to keep the lights on upstairs. Manual-automation, manumatio…

Scatter guns and muskets.

Many, Many years ago I worked at a startup called (a European online travel company, back when a travel company didn't have to be online). For a while, I worked in what would now be described as a 'DevOps' team. A group of technical people with both programming and operational skills.

I was in a hybrid development/operations role, where I spent my time investigating and remedying production issues using my development, investigative and still nascent testing skills. It was a hectic job working long hours away from home. Finding myself overloaded with work, I quickly learned to be a little ruthless with my time when trying to figure out what was broken and what needed to be fixed.
One skill I picked up, was being able to distinguish whether I was researching a bug or trying to find a new bug. When researching, I would be changing one thing or removing something (etc) and seeing if that made the issue better or worse. When looking for bugs, I'd be casting…