Nerdherding for Beginners

Sometimes nerds sneak in unexpected features to the things they create. I’m not talking about easter-eggs but features they believe will be of value to the product’s users. This is a bad thing to do which is often a surprise to the offending nerds. Its good to be proactive and to offer additional low-cost features to customers but adding them without checking first is imprudent.

Most of the time the nerd who inserts the unexpected feature isn’t the customer (or the customer’s representative). So why is it okay to spend someone else’s money on something they might not want? Maybe they would have wanted something else instead. Maybe they would have preferred the software to cost a bit less.

There is also an additional cost for every feature exposed to those who use software. Someone needs to explain the feature, it needs to be tested, maintained, and so on. So even if the feature truly didn’t cost anything to implement (say it was autogenerated by an existing application framework) then someone’s money has still been spent without checking first.

GPSports would like to hire a Java nerd for a few months to help us out with an impressive looking deadline. If you’re looking for a fun job in a neat company in Canberra for the next few months then drop me a line.

Sometimes being around computers too much can send a nerd “binary”. They begin see things in extremes - all or nothing, something is done or ignored. For example, suppose that documentation of a proposed feature is needed for a team that has never produced documentation before. I’d probably start small and work up to a reasonable level by over a number of iterations. Those nerds that have gone binary would immediately jump to the rolls-royce solution and create (or propose) a fifty page monolith complete with sign-off forms and paragraph level revision numbers.

This problem manifests itself in other ways. Need to write to a log file during installation? Instead of using Jakarta Commons Logging or writing a simple class to do it, they feel the need to design and implement something that approaches the functionality in Commons Logging but obviously takes much longer.

What can be done about it? Not much apart from constantly being on the look out for signs it is about to emerge, and learning which of your nerds have gone binary so you can keep a really close eye on them. Those who have gone binary seem to be generally unaware of it and are often unable to catch themselves unless they are about to overdo the same thing again (ie writing documentation).

Sometimes people are ready to go and solve big nasty problems or lead a team for the first time or design a custom memory manager for that C++ server application. Sometimes not. The trick for managers is to be reasonably sure that your chosen one will succeed. If you aren’t sure then put in a safety net or someone is likely to get hurt. Consider providing a mentor/coach for the task. Another idea is to meet regularly to track progress and to arrest the slide into failure or to mitigate the consequences.

Managers have a duty to ensure that their staff are not placed in a position where they are likely to fail or where failure will cause them to suffer. To do so unfairly risks reputations, happiness and health. I’ve seen too many decent, hard-working people damaged by careless and selfish task assignments. This is not to say that people should be mollycoddled. There is a middle ground.

Do you send people out to fight dragons before they are ready?