Classic Software Mistakes: Steve McConnell

Steve McConnell's blog is advertising a survey on the prevalence and severity of software management mistakes. It's worth taking it, just to see the long list of possible disasters. I don't think it's a wonderful survey as surveys go (they are very hard to do well), but it might make you feel slightly better about your current development environment when you see all the things that can go wrong. Or, not.

Some examples:

• Insufficient planning
• Abandonment of planning under pressure
• Subcontractor mismanagement
• Lack of effective sponsorship (at executive level)
• Outsourcing to reduce costs
• Unclear project vision
• Switching development tools in the middle of a project
• Lack of automated source code control

They're all defined during the survey, so it may feel a little long to take (plan on 20 minutes). My biggest difficulty was that a lot of them overlap, and it's hard to tell which one to "vote" for! I also wondered what he meant by "catastrophic" results -- I decided it meant people quitting, rather than projects getting killed, since it's often hard to consider the latter a catastrophe instead of a mercy killing.

I admit to mixed feelings about McConnell and this survey doesn't help reduce them -- he and his books are written as if no one but developers are involved in building software, which alienates me as a designer. His discussions of requirements, planning, and design don't seem to allow for the involvement of user researchers or even product managers, let alone interface designers who may have to write the specs. One "classic mistake" I see is to leave all the interaction design, visual design, requirements collection, and user testing to developers who also need to be writing code; may not know how to do the non-coding work well; and may not want to do it "all" themselves.

Software is and has been a multi-disciplinary effort for a long time now. McConnell has a huge readership, but hasn't helped spread that word, which set us all back a bit, I think.