Avoiding IT screw-ups is simpler than you think

You usually don't have to wait too long before some sort of IT disaster appears in the news headlines: the thing about IT is, it's complicated. Lots of different things have to work together in tandem, otherwise cracks start to appear — that could mean anything from the occasional bug to the whole system coming crashing down like a house of cards.

That's why there's a running private joke in software development circles, known as Works on My Machine. The comedy behind the quip is that just because code runs on one system doesn't necessarily mean it's going to run successfully everywhere else — or that other users are seeing the same results as you are. Again, software relies on lots of ducks being in a row in order to work: the most careless developers sometimes release new code without a thought to all the other factors that their programming relies on.

Updating to a new software version, changing a configuration setting, or introducing a new feature can be the root causes of all sorts of issues, and these are just a few of them. The broad types of problems listed in that link — from the bad weather that took down Amazon's servers to a number of overflow bugs that crashed a space rocket — should give you some idea of just how delicate software development is, and how carefully programs (and the systems that they run on) need to be put together.

IT failure is not an inevitability though, and there are ways to address and minimise the risk. One of them is through the use of automated deployment, an idea we're fully behind at D4. In fact, the more automation the better, as it means less opportunity for human error.

Another way to prevent future problems is to borrow an idea from aviation: the humble checklist. The last time you were on a flight you might well have heard the pilot say something like "cross check and doors to manual" — that's because the plane's crew are carefully following a checklist to make sure nothing is forgotten or overlooked.

It's something that those in the world of medicine are discovering has its benefits, and if it's good enough for the World Health Organisation, it's good enough for us. It scales well too: whether you're trying to bake a cake or organise a group skiing holiday, a checklist can prove invaluable. Some of those we find useful at different points are the new starter checklist, the requirements-gathering checklist, the deployment checklist and the project kickoff checklist.

So, IT screw-ups — they happen easily, but they're not impossible to avoid. If you want to put us to the test, get in touch.