Quit making my testers dumb!
When I see some of the tools available to "aid" testers, I alternate between shock and anger and disappointment! Don't get me wrong - there are some wonderful tools out there; but soooo many tools seem to make an assumption that the tester is more or less a moron who cannot point out an error when he sees one! I am sure many folks resonate with this crib - here is one.
The last discussion that we had about video recording is in my mind generated a conflict in my mind as well. As an oversimplification, let's divide testers into 2 categories: those that debug and those that won't. If I offer a feature like "video-recording-inserted-in-bug" to the former kind, on its own, it's almost always a waste. When I see an issue that does not repro, I try to bunch all invariants and eliminate the changing factors one by one to arrive at the culprit factor that is triggering the bug. In that case, a .wmv file by itself is useless. I'll need data like proc usage, mem usage, env vars set at the time etc. to arrive at the answer. Instead of doing that, if I just attach a video, it hardly helps the dev either - no one benefits.
But consider the latter category: these maybe domain experts, tech support, beta testers, bug bashers - basically a set that may have debugging skills, but are not willing to use them in this situation or a set that do not have debugging skills at all. In that case, attaching a video file may be helpful to recount parts of the bug tale that the testers may not have observed. The video clubbed with system information captured automatically as well as event logs form a powerful set of debugging aids for the developer.
But, abuse of these features can often lead to tester skills being underdeveloped. For instance, if a tester attaches the video and says "Okay - here is the video. I dunno if it were the mouse middle click or the multi monitor or the fact that my CPU was 90% busy that caused the crash; but hey, use the video and go figure." That is a pretty stupid thing to do. Essentially, the feature ends up making the tester hesitant to narrow repro steps, debug the issue or even think up similar cases. You are making my testers dumb!! Or worse, lazy! Thankfully, this is a minor problem where just regulating the QA processes comes in handy. From what Bj said in the last post, a product tester would probably have debugged more into the issue, tried real world testing and found the cause of the bug. But as Chris and Sumod point out, perhaps the repro is coming in from people that are not interested in/don't have time for debugging into this issue at all.
With testers coming from such a wide range of background/experience and requirements, I find the current testing tools available in the market to be woefully inadequate as well as ill targeted (both in the automation and manual testing space). Do write to me about reqs that you have had a tester that have not been addressed by any testing tool at all - perhaps we can collectively devise a solution for our common problems.
And as a postscript, please don't take the title too literally and flame me and accuse me for calling testers that do not want to debug as "dumb" - that's certainly not my intent.
