This week's editorial is a guest post from Phil Factor.

I once managed a software house that did applications for several local businesses. It wasn't easy, because these businesses ran on a very tight budget, and they didn't want the frills. We worked on fixed-price contracts, so if we got it wrong we made a loss. It is a great way of developing a sense of urgency, and making sure that your programming work was sufficient but never fancy.

Our database guy was a quiet, methodical chap who went about his work slowly and carefully. He took every deployment seriously, even if he was making a change to the custom music player in the server room, cobbled together from an ancient Unix box.

As a manager, I gradually began to feel resentful that all his database deployments were taken so seriously, and took so long, even when it didn't matter, and that our development time was getting squeezed in consequence. Eventually, I challenged him about it. To my surprise, he came up with a very good reason.

The reason that every database deployment is taken with the same caution and seriousness, he explained, was that they generally went well; to the point that he could easily become rusty on the many roll-back procedures, belt-and-braces precautions and tests. When anything failed, he tested his skill in bringing things to a happy conclusion. We were managing applications where failure would cause businesses and their staff real catastrophe. A deployment gone wrong could, unless quickly remedied, cause significant loss of trading or reputation to these guys. Because our database guy treated every deployment with the same attention to detail, he occasionally had failed deployments to deal with that weren't important to the outside world, but allowed him to fine-tune his deployment scripts, procedures, and techniques.

I bought this explanation. We never ever had to apologise to a customer. It reminded me of the craftsman restoring a cathedral who maintained his meticulous work, including lively and beautiful carvings, in parts of the building that could never be seen close-up. When challenged to explain why he took such care over those parts of his work that nobody would ever see, he merely replied enigmatically "Well, God will see it." Of course: what they were driving at was that by honing your skills where mistakes don't matter, it allows you a better chance of getting it right when it does matter.