“He’s mad that trusts in the tameness of a wolf” so said The Bard’s Fool.
Just as mad is he who trusts the tameness of a software project.
Building or implementing software at any significant scale always involves lots of humans interacting. Whenever there are lots of humans butting heads, especially on something as complex as building new software machines, all we can really be certain of is that results will be uncertain and that behaviours will be anything but tame.
So, in this context, I find myself close to yet another large-scale Waterfall-based software implementation. And yes, that’s right, it is 2014, thirteen years after Beck, Fowler, Cockburn and friends made their revolutionary pact in the snowy peaks of northern Utah.
Here’s how it goes on my current project:
- Create an invalid plan
- Realise after 2-3 weeks that the plan is invalid
- Deny that the plan is invalid
- Wait until the latest viable moment to admit that the plan is invalid
- Change the plan and/or the project managers
- Rinse and repeat
Me: “This is madness. We create these plans that we know are not worth the paper they’re written on. We get everybody ‘signed up’ to them, then become completely disillusioned when they turn out to be bollocks. All this pain and angst gets wasted every time we do this.”
Client: “Yes, yes, but we have to have a plan. How can we possibly manage a project without a proper plan?”
Me: “OK,” I attempt to compromise, “how about we create a plan just for the next 4-6 weeks, and have some basic milestones for further out.”
Client: “That doesn’t work because we need to align resources for further in down the line.”
Client: “What, align them against a plan that we have 10% confidence level in?”
Client eyes glaze over: “Well, we have to a plan that we have confidence in.”
Me: does anyone have a good fork handy for me to remove my prefrontal cortex, so I don’t have to comprehend any of this anymore?
What I’m describing here is the impacts of what Daniel Kahneman describes as the planning fallacy. Whether it’s engaging in wishful thinking or overly discounting past failures or off-project risks, we humans don’t do this type of planning well. That’s one of the reasons much of the industry is now ploughing the Agile furrow having effectively given up with long-term detailed plans.
Agile of course has its own sets of flaws and thinkers and writers such as Dan North and Jim Bensonare now providing fresh critiques on some of our established Agile practices. However, none of these ‘post-Agile’ perspectives envisions a return to Waterfall. Please let’s stop with the silliness and accept our human limitations. Gantt charts have their place, but when used to predict outcomes for complex projects, we’d be better throwing them to the wolves.