I was spending some time with a software test engineer (tester) recently and have a simple realization to share.
“How come this runs so slow?”
“Oh it runs just fine, just some problems only in IE browser”
“Yes but soooo slow?!?”
“Yes… but my Lead has seen it and knows about it”
“Which means it doesn’t matter?”
“Well not that… client has also agreed to Chrome-only support”
“The site will support only Chrome?”
“Yes, we described the problem and got client to agree to it”
“Do you have it in writing?”
“Very good. But why is this the case in first place?”
“Developers told me the 3rd-party component doesn’t run on IE.” (look on face: how does it matter since client has agreed anyway)
“You mean a best-selling web UI component library doesn’t run well on IE???”
“Yes that’s what developers told me”
After some pushing and research, it finally turned out that developers had to use another method of data access and voila! – everything ran smooth on all browsers.
No doubt, the incident above and its subsequent resolution display several best practices. Get acceptance criteria in writing. Make sure client understands limitations. Don’t trust explanations! Be hard-nosed. Use common sense (how can a best-selling web UI library possibly not function well on IE!). Don’t let your personal love override the obvious (Chrome is great but that doesn’t mean you can force users to switch).
An aspect I would like to focus on is: to whom should the tester be responsible? Project Manager? Team? CEO? Entity/person who is paying for work?
In the course of project teams and organization and “reporting structure” and salary and revenue and a zillion other things, we forget what is obvious. In reality, the Lead/Manager doesn’t matter. The developers do not matter. The CEO doesn’t matter. The client doesn’t matter! They are all illusion, maya.
The only person who matters, the only person who finally decides, is the user. A software tester is responsible towards only the system being tested, and users of that system.
In the short run, you can get away by keeping your Lead/Manager happy. You can get away by accepting superficial wave-of-hand explanations from developers. You can get away by getting clients to agree to something illogical. You can get away by using bureaucracy to cover your behind – “I have it in writing”.
You can get away because the short run is imperfect. The short run can hide. But the long run is all about the right thing. The good thing. The correct thing. Quality. Truth. Details. Finesse. Care. The long run exposes everything.
The right thing here is to find and fix the problem, or find a workaround if it cannot be fixed.
If you do not, all might be well at the surface. But when the user rejects it in the marketplace, everyone loses. And why? Because we ignored her when we should have paid attention. We got confused by irrelevant intermediaries (colleagues, lead, client, CEO) and forgot who the software is being built for.
In the example above, the user is taken for granted. The user is actually expected to tolerate slowness, or switch her browser if she wants to enjoy our system. The IE user votes No and many users are IE-only. Yikes, that pinched!
As a software tester (whether for a web site or web application or desktop system or mobile app or server software and so on), you are the user’s voice and warrior. Don’t accept reasons. Make your point. File issues where you see issues. Revolt till someone notices. Help in getting the right thing done. Ride alone if you have to, but ride the right path.
Honestly, you are not alone in getting deceived by quick things. It is the same thing in business. Business gets confused by targets and profits and “customer satisfaction metrics”, because those are the quick and measurable things, although they are not what business is about. Business is about doing the right thing for people, not only the legally right thing.
Software testing is about doing the right thing for users, not only the officially right thing.
Author – Chinmoy Panda