Thoughts on Agile
by Gordon Weakliem
I have thoughts on Agile. Yes, I do.
I was introduced with the original Extreme Programming Explained, somewhere around 1999. It seemed like a great idea. I was obsessed for a while. The problem is nobody on my team liked it, particularly the part about pair programming. Around 15 years later, I got to work on a team that practiced what seemed like XP, they didn’t identify that way, but their practices were very much aligned. It was exhausting, and exhilirating. We got a lot of work done. One thing I noticed is that we were always outrunning our PMs. A team of 8 was constantly out of work and our 2 PMs were constantly scrambling.
That’s how it is now. If you’re doing it right, whoever owns the project is scrambling to keep up. It may not seem like it, but they are. You don’t even have to be doing it right, just a little boost, and your idea people run out of ideas pretty quickly. Or the ideas are terrible and by the end of a sprint, they deliver ideas that are not that important.
Ideas are cheap, programming is hard, but now that changed.
I want to get back to Agile. Why did we need Agile around the turn of the century?
- Because ideas are cheap, but time is expensive. People will intrude all the time with things that seem important. They might be important, but triage is a big issue. In an uncontrolled environment, triage becomes all you do. Part of the point was to separate the implemenation from the idea: let it bake a little to see if it’s really worth doing.
- Because communication is expensive. Communicating a vision well is hard, we call the people who do it well “visionaries”. Communicating a simple idea is hard. In an engineer’s context it’s hard because things fail in a lot of ways. It’s also hard because the engineers don’t understand the vision, and vision is hard to come by.
I think the value in XP is to explain things in simple terms. The idea I took away is that nobody pays for plans, they pay for results. This plays out now because so many barriers are gone. The barrier is you, explaining what you want.