Software Engineering Design Decisions - Three Bad Practices

BY MARKUS SPRUNCK

Maybe you know the joke where a young software engineer goes into a bar, puts a green frog on top of the bar counter and the frog says: "Kiss me, I'm an enchanted princess." The bar keeper is fascinated and recommends the software engineer to kiss the frog. But he just replies "I have no time for a girlfriend and a talking frog is cool!".

We love cool things, difficult technical puzzles and sometimes challenging bugs. But we should be careful with our motivation if we do technical decisions. I have seen and will see a lot of stupid decisions from engineers and/or managers. The following list of three bad practices is neither complete nor scientifically verified, but I think you may recognize some of them in your organization.

Big-Mac-Menu-Design-Decision

A Big-Mac is a product with consistent standard of quality in space and time over the past 40 years. Just small variations in nutritional values between countries [1] and almost no variation within a country. But this doesn't says anything about taste and/or healthiness.

Large organizations tend to have a enterprise architecture department which produces a kind of standard architecture. Independent of the selected technology a standard architecture has usually a lot of overhead (because it should fit for every project in the company) and/or it is outdated before published for the first time.

It can be dangerous to use a standard architecture in inappropriate circumstances and it is difficult to do decide for alternative solutions.

Please, try to courageously raise your voice if the standard architecture doesn't fit for your task. The enterprise architecture department should work for us and not vice versa.

Swarm-Foolishness-Design-Decision

Often, a group of humans is not more intelligent than each single team member. I know that this is a provocative statement, but it's a matter of fact that team dynamic can be a killer of intelligent decisions.

If a group of people discuss a problem and every body's input is told, it is very likely that the person will get acceptance of the team which is most sure that he/she knows the right solution. People which are convinced that their opinion is the correct solution are so dominant to others, that even people which know it better getting uncertain and just recede from their opinion.

The only way to avoid this pitfall is to believe nothing. Try to find a real prove, a sample, some measurement, program spike solutions and/or reliable case studies. Software Engineering is a discipline that should be founded on knowledge and not on believe.

Manager-Design-Decision

Unfortunately relatively often these people which are perfectly sure that they have the right solution are managers.

A lot of organizations promote people which are self-confident and good in solving problems. Over the time these managers lose their technical competence, they just stop to be up-to-date. Some are still convinced that they know everything better than their team members. This can be a great problem, in cases where the manager is fooled into believing that a manager is automatically a good designer.

In the case that your manager permanently over rules the team and makes stupid design decisions, maybe it's a good idea to find a better manager. Or you could print this article, use a yellow text marker to highlight right paragraph and drop it at the desk of your boss. At least you will have some fun and there is a little chance that he/she will improve.

References

[1] Big Mac Wikipedia; The Big Mac is a hamburger sold by McDonald's, an international fast food restaurant chain.