Having benefited from the imagined history of Object-Oriented Programming, it’s time to turn our flawed retelling toolset to Agile. This history is as inaccurate and biased as it is illuminating.
In the beginning, there was no software. This was considered to be a crisis, because there had been computers since at least the 1940s but nearly half a century later nobody had written any software that could run on them. It had all been cancelled due to cost overruns, schedule overruns, poor quality, or all three.
This was embarrassing for Emperor Carhoare I, whose mighty imperial domain was known as Software Engineering. What was most embarrassing was that at every time when the crisis came to a head, a hologram of Dijkstra would appear and repeat some pre-recorded variation of “I told you so, nobody is clever enough to write any software”.
In frustration, software managers marched their developers off of waterfalls to their doom. If you ever see a software product with a copyright date before, say, 2001, it is a work of fiction, placed by The Great Cunningham to test our faith. I know this because a very expensive consultant told me that it is so.
Eventually the situation got so bad that some people decided to do something about it. They went on a skiing holiday, which was preferable to the original suggestion that they do something about this software problem. But eventually they ended up talking about software anyway, and it turned out that two of them actually had managed to get a software project going. Their approach was extreme: they wrote the software instead of producing interim reports about how little software had been written.
With nothing else to show than a few photographs of mountains, the rest of the skiing group wrote up a little document saying that maybe everybody else writing software might want to try just writing the software, instead of writing reports about how little software had been written. This was explosive. People just couldn’t believe that the solution to writing software was this easy. It must be a trick. They turned to the Dijkstra projection for guidance, but it never manifested again. Maybe he had failed to foresee this crisis? Maybe The Blessed Cunningham was a mule who existed outside psychohistory?
There were two major problems with this “just fucking do it” approach to writing software. The first problem was that it left no breathing room for managers to commission, schedule, and ignore reports on how little software was getting written. Some people got together and wrote the Project Managers’ Declaration of Interdependence. This document uses a lot of words to say “we are totally cool and we get this whole Agile thing you’re talking about, and if you let us onto your projects to commission status reports and track deliverables we’ll definitely be able to pay our bills”.
The second problem, related to the first, is that there wasn’t anything to sell. How can you Agile up your software tools if the point is that tools aren’t as important as you thought? How can you sell books on how important this little nuance at the edge of Agile is, if the whole idea fits on a postcard?
Enter certification. We care more about the people than the process, and if you pay for our training and our CPDs you can prove to everybody that you’ve understood the process for not caring about process. Obviously this is training and certification for the aforementioned co-dependent—sorry, interdependent—project managers. There is certification for developers, but this stuff works best if they’re not actually organised so you won’t find many developers with the certification. Way better to let them divide themselves over which language is best, or which text editor, or which blank space character.
And…that’s it. We’re up to date. No particularly fresh theoretical insight in two decades, we just have a lot of developers treated as fungible velocity sources on projects managed top-down to indirect metrics and reports. Seems like we could benefit from some agility.