After I mentioned our bug problems at work, Andy ironically suggested we think outside the box.

Nope. Turns out we had the opposite problem: full story courtesy of our flown-out-to-site victim Dave Miller.

It's great to have the problem "fixed", but it's soooo irritating - I remember several kitchen conversations along the lines of "the irreproduceability means it's tempting to attribute the crash to a hardware error, but that's just wishful thinking: <sarcasm>Yes, coding team, of course there's no problem with the huge amounts of new code you wrote, it has to be someone else's fault...</sarcasm>". And if it had been easier to get anyone at the test site to run some diagnostics, it would have happened. But probabilities / coincidences, plus the lack of access to and remoteness of the site made us judge it to be not one of the best things to spend our time on.

Funnily enough, this has happened before. When a bunch of us visited our Japanese reseller a few years back, they complained of an intermittent crash in our software that other Voxar Dave (Turner) removed by opening the computer and snapping the motherboard more securely back down into place...

Tags: software_development
Current Mood: rueful