Sunday 25 April 2010

System Crashes. Also needs ellipsis

A recent post on the STC from Tony Simms asking for Real Examples of Bad Bug reports got me thinking.

Sadly, I could have supplied him with a page full of examples.

Going back to my days as a dev, the company I was working for had one 'tester' and he had the most aggravating bug reports.

"Selecting foo from the menu bar crashes the system. Also foo needs to end with an ellipsis"

So I'd fix the crash, forget to add the ellipsis to the menu item and the bug fix would get rejected. I had some petty revenge when I was put in charge of testing and would reject bugs that had more than one defect reported.

This was several years ago but I still come across poorly reported bugs ( including ones similar to the above where two defects are reported in one report )

One solution seems to be to use the approach that MS outlined in their demo to the London Tester Gathering of VS 2010. All the tester has to do is click a button and the defect and its environment are sent to the developer.

Or the testers can learn Bug Advocacy. Check the date on this famous paper - 2002 and still relevant today ( and still unread by a lot of testers I've worked with )

Maybe it's part of the dumbing down of testers - give them a tool to use rather than teach them how to report bugs.

Eric Jacobson recently posted about The Joy Of Cracking Repro Steps - another skill that some testers don't seem to have. Find a bug, log the bug - no effort to find the simplest way to reproduce the defect or exactly what causes it. My recent example of a Campfire Bug is a case in point - I got a great buzz out of narrowing it down to having spaces in the mobile phone number field and it meant that when the dev had to fix the defect all the information required was there.

but if you are only measured on how many tests you run and how many defects you find, what incentive is there to narrow down a bug and report it effectively ?

No comments: