Last week, I had the priviledge of attending the Saturn 2013 conference in Minneapolis. Those who follow @eltjopoort on Twitter may have already seen some posts about it, but I thought I’d summarize the hightlights in this blog. SATURN is an important international architecture conference, organized annually by the Software Engineering Institute. It attracts an international audience of senior architecture professionals and thought leaders from all industries. All slides and summaries of the talks can be found on the SATURN conference website.
The first tutorial I attended was “Coaching architects” by Felix Bachmann on Monday. Felix talked about non-technical aspects of coaching or leading a group of architects, designers, lead-engineers etc. in a project. For me, the key take-away was his tip to “speed up the process of making mistakes” if you find people are slow to pick up your coaching. That way, people get quicker feedback and learn faster. Felix also repeatedly confirmed the RCDA principle that a leading or coaching architect is responsible for the architecture and the decisions that shape it, “even if the organisation says differently”.
On Tuesday, I gave a tutorial on RCDA’s principle of prioritizing architectural concerns by risk and cost. There was a lively interaction, and positive feedback from the attendants.
Up until Wednesday, we had had nice spring weather at >20 Celsius: then a rare May snow storm hit Minneapolis… fortunately, the Minnesotans had no trouble whatsoever dealing with that. The first keynote was by Stephan Murer, IT director at Credit Suisse. He presented the lessons learned of 15 years of SOA. Very interesting to see the lifecycle of this hyped technology summarized so nicely. Startling fact: it took 5 years (1998-2003) before the SOA services (CORBA-based) started to be used by more than a handful of applications – up till then, the business case for SOA at Credit Suisse was based on small gains from individual projects. Nowadays (2013), the application landscape completely depends on the SOA services exposing the company’s data, and the architecture has paid itself back multiple times. This story shows that long-term decision making is key in enterprise architecture. It also stresses the importance of a vision which may be based largely on belief in good design rather than demonstrable ROI. On the other hand, such a vision can only be sold to management if it is presented with economic arguments, or “you may have a short life as an architect”… Wednesday evening, I gave a lightning talk on assumption-free architecting, which according to the response from the audience was either quite funny or laughable… see this previous post.
On Thursday, Simon Brown was the winner of the Architecture in Practice presentation award with his talk about the conflict between agile and architecture: myth or reality? (short answer: it’s a myth). Simon is a great presenter, and his award well-deserved. I’ll come back to my own presentation on lean and mean architecting in a later blog.
The closing keynote by Philippe Kruchten was very entertaining. Philippe discussed cognitive bias in architectural decision making; this seems to become an important theme in the architecture community. Philippe’s basis was Kahnemanns “Thinking – Fast and Slow”, a book about which I blogged earlier. Architects have to deal with a lot of cognitive bias, either unconsciously or consciously induced in reasoning fallacies and political games. He gave a long list of (often amusing) examples of how architects can make wrong decisions due to anchoring, confirmatory bias and availability bias. We have all run into “Golden Hammer” (everything looks like a nail), “Elephant in the Room”, “Not Invented Here”, “Analysis Paralysis” and “Perfection or Bust”. Organizations often engage in architectural “Cargo Cult”, imitating the externally visible behaviour or architecture that appeared to work in another organization, hoping to achieve the same results – without verifying applicability of the underlying mechanisms in your context. Let’s kick our System 2 into gear!
On Friday, I attended two tutorials. Mary Poppendieck (LEAN high-priestess) taught us the LEAN mindset in a nutshell. I was amazed when she very simply demonstrated that optimizing a delivery organization for utilization inexorably leads to a substantial reduction of throughput! This effect from queueing theory starts at utilization over 50% if the batches of work are large. Powerful illustration: “Highways that are fully utilized are called parking lots”. I need to talk to management about our KPIs… J.
Finally, I attended Philippe Kruchten’s tutorial about roadmapping. RCDA’s Architectural Roadmapping practice is based on Philippe’s work, so there were no new insights for me – but we did have a lot of fun playing his Mission to Mars game.
All in all, a very enjoyable conference –presented RCDA, made new friends and learned a lot. Many thanks to all who made this possible.