Agile on the Beach (AotB) is a two day conference and really packs a punch.
With a three day conference you can feel yourself settling into a community and find a rhythm at the end of the first day and go from there. AotB just hits you in the face from the start and flies by just like that — short and sharp.
Part of the experience is spending most of a day travelling. I’m sure part of it is that if it takes more effort you appreciate it all the more; that and the shear quality of the line up, especially on the “Technical” track, or Software Delivery as it’s titled.
It is highly competitive to get in to to speak, and I’m extremely proud to say I was part of that line up this year; but it took a few attempts, and finding the right topic. Partly this is because of only being two day.
With Steve Smith, a UK practitioner authority on Continual Delivery in my book, and Daniel Bryant, always providing me with the Zeitgeist of container delivery, on each year I’ve been, the standard is high. This year Simon Brown of the C4 model was also speaking, and the tools he’s developing have a lot going for them.
Then there’s Gwen Diagram adding some color, both in terms of her hair, and her language. Two years ago she stole the show at the Lightning talks, then last year as Endnote and this year as a speaker. Always a rant, always a bit sweary. But her enthusiasm is infectious.
Over the three years we’ve had some stellar names, a James Lewis-Dan North double act, James Grenning, Kevlin Henney, Woody Zuill; which led to us to do some pair programming and then visit Redgate a few weeks ago. There’s also been fascinating talks on dealing with uncertainty and last year the rollercoaster of Gwen’s Endnote. And all the time Allan Kelly doing his own slots, and sometimes others too.
Allan always says he try to make each track worthy of a standalone conference in it’s own right. And I think he does.
Last year I traveled down with two friends and was introduced to the game Fluxx; which I then bought a couple of members of the family for Christmas and Birthdays. I was also nearing the end of the Harry Potter series (for the first time, as it happens).
This year I was travelling by myself, but then I had my slides to work on, so the journey relatively flew by.
So having survived the pasty, pint and pub quiz on the Wednesday night, and a glorious freshly cooked egg, bacon and sausage muffin, onto the opening keynote.
Dr Tendayi Viki spoke about how Innovation is not a linear process, and squeezing it in to 4 weeks of one phase and then 12 of the next makes no sense; just like doing Waterfall Software Development.
After the break, I was split between seeing Ceri Newton’s talk as I missed it at Agile Manchester; but I was also conscious I had to see Simon Brown’s talk on C4 Architecture. I tried to go to Ceri’s but it was full and ended up in a live-stream room watching it; but then thought I may as well got to a live talk and catch Ceri’s on video afterwards.
Simon was on in the main hall, which the delivery/technical track gets all the way through; which is a nice declaration of priorities of the conference, so I had no problem getting in.
Simon was going through the 4 layers of abstraction, or Cs, and was explaining the importance of labeling diagrams and components on them; use colors to signify different types; provide a key on each diagram for ease of reading. The whole point is to communicate effectively and concisely your architecture.
He emphasized the 4th, lowest, level of detail should not be drawn manually, but produced by an automated tool such as an IDE.
I did a C4 workshop a few years ago, coincidentally at an evening meetup at Redgate. The talk made me think about the usefulness of these diagrams in helping developers explore, understand and communicate about the architecture. As well as see the bigger picture and context in which they’re working. I’m planning to kick off some of these workshops to up Redgates Architecture game.
I value the tool as a way for a developer to explain what they’re working on, not keeping the diagrams as artifacts. But as I understand it, Simon works with large Financial Enterprises who like to retain an element of “control" over decisions.
Next up was Steve Smith. I enjoy Steve’s talks for his dry humor and how he outlines things you really should be doing. As it happens Australia were playing England in the Cricket World Cup Semi-final during the day, and were 14–3 about the time of the start of the talk; Steve started off by baiting any Australians in the audience. Notably, during the course of his talk, Australia lost no more wickets (but did go on to lose).
Steve described how he started one assignment and aimed to get the lead time for delivery down from 34 to 14 days. He managed to reduce it to 31 days; but in hindsight found that there was a 15 day delay he could have tackled if he’d asked the right questions. He persuaded a Product Owner to use a 2 over a 1 week cadence as there was a 9–11 day process that wouldn’t fit in 1 week. This created a stable feedback loop for the iterations.
Steve took us through the Theory of Constraints (described in Goldratt’s The Goal) which says that tackling anything but the system’s bottleneck is a waste of time (and resources, and people for that matter). Once you’ve tackled a bottleneck, then tackle the next to emerge.
I recommend Steve’s talks, I’ve seen a few live now, and some online. I think it was Steve’s at my first AotB that inspired me to go ahead and setup Continuous Delivery on a project within a week of getting back that year.
During lunch I saw part of a talk on Creative Coding for Kids; this was a nice touch by the organizers.
For the next session, I caught part of Turning the Oil Tanker, on introducing Agile to Cornwall Council over the last year, after a few employees attended last year. I’d spoken to Suzanne Manson over lunch and it sounded interesting; though again I was in a Live-Stream room.
Eventually I decided to catch the end of Gwen Diagram’s talk as there’s always fireworks, but only really caught questions at the end.
For the final session there was a session called Real World Iterative Database Development, which was so close Professionally I had to attend; it was also co-presented by Chris Pitts who I knew personally from previous years.
The session explored normalizing data, but iteratively and according to need. There was a hands on exercise, but it took too much of the session to get Docker working on my laptop; but the talking and discussion made the session worthwhile anyway.
Ron Ballard, the other co-presenter, gave his views on a few things, and interestingly said to avoid database Views. I spoke to him outside of the session as I wanted to understand his thinking, and as I suspected it partly came down to an ossification argument — once you’d created one, it was hard to get rid of —they’re like publishing an API which you then need to maintain. He also made the point that one is fine, but you soon end up with 5, and three that consume each other, then you’re running 3 queries, returning lots of data to get a single piece of information.
The Thursday evening is always the Beach Party, which involves a barbacue and bar, with cask ale; some sea shanties, an ice-cream cart, toasted marshmallows on braziers and running for buses back to campus. It’s great to chat and meet new people, and catch up with some familiar faces; even in 24hrs we’d developed some repeating jokes.
As I was speaking the next day I took the first bus back to review my slides, and avoid drinking too much…
…to be continued. I’ll describe day two in the next post.