Santa Claus & Bell Ringing — Redgate’s Open Space #2
We decided to get a second Open Space event in before Christmas, as we try to turn these events into a regular way for the company to come together to learn and make productive use of our 10% time.
Based one feedback we had tweaked the format, shorter, more composable sessions with a break half way through the event to try to give the flexibility that comes with an Open Space.
Personally, I had hoped for more topics, but the topics that we did have were interesting and the sessions were interesting.
There is no Santa Claus… or the Spotify Model
Ben Mancini gave us a report from when he was at the Lean Agile Scotland conference where he discovered that Santa (or the Spotify Model) does not exist. For many years, Spotify has been held as a beacon for the Agile community, they were seen as the forerunners, the thought leaders in this space.
The current reality is far different, due to the pressures of hyper-growth they have abandoned many of their ideas. Look around the offices and you see people in headphones, not collaborating/pairing, you see blank whiteboards/dashboards and there has been a massive growth in management roles.
We care about releasing frequently, we want to move faster, but Spotify did not come up with a Model for us all to copy, they went through a process to reach a set of ideas that worked for them at the time.
I think one of my biggest takeaways was a quote attributed to Taiichi Ohno (the founder of the Toyota Production System): Stop trying to borrow wisdom and think for yourself.
Rather than blindly copying an existing process, we should understand the tools available and face our own problems.
Enums are “evil” vs the Open-Closed principle
This was something that Jeff Foster proposed and has written up in his own post about the Open Space.
His principle statement came from working on some of our shared libraries that contain enums and finding switch statements dotted across our products. There were many people who preferred the type safety that comes with enums, but in C# there is no enforcement that you have handled every case, and there is coupling so that should I delete a value I have to update all clients for them to compile (and potentially communicate).
We ended the discussion with more of a belief that this is more about encapsulating your dependencies behind an anti-corruption layer.
Why does everything take so long?
A common feeling, probably among many people in a company, Gareth Bragg decided to ask this question (despite telling me barely 30 minutes earlier that he had no ideas to put forward as sessions this time). It turned into a long discussion, that ran through the break and into the next session (it ends when it ends) with a lot of input. Gareth did a fantastic job of mind-mapping the entire conversation.
A few things really sprung out at me:
- Perfectionism — we try to get things perfect first time, we think we can, and that may stem from a fear of failure.
- Consensus — we try to get everyone to come on-board with a decision, we try to reach a consensus with everyone, which often leads to bikeshedding.
- Trust — some of the difficultly in reaching consensus may also come from a lack of trust, for example who really believes “we’ll revisit this once we’ve shipped?”
One thing I suggested is that we look at Real Options which could help with decision making. The basic premise is that we do enough work now to know what our options are, and when those options expire, then we can delay the decision to exercise the option at the appropriate moment. Most of the time, we can debate for that long that most of the options expire and we are left with the ones we liked the least.
Sort Algorithms With Bells
At the end of the Open Space, I ran a session involving some bells and sorting algorithms. It was inspired after attending a session run by Matthew Butt at the SoCraTes UK Conference this year.
With just 8 bells (an octave) can we reproduce and compare different sorting algorithms? (*)
Each person takes a bell and stands in a line, we take it in turns to compare against another person to determine if you are a higher, or lower, note and sort yourself accordingly.
It is a fun session that gets people moving and interacting, it is a good way to teach, or reinforce the differences between the algorithms and it leaves me thinking where could we go from here? Bit manipulation of a number using an octave of bells?
* It can get a bit loud, so be careful if you want to try this out.
