Live Blog – Dave Farley – Continuous Development (Keynote) 2015

Dave Farley is the author of ‘Continuous Delivery’ and the ‘Reactive Manifesto’. He has 30 years experience of software development and was an early adopter of agile development techniques. You can catch up with Dave at his website and on Twitter.


A long time ago in a canteen far away…….

Dave dreamed up Farley’s three laws.

  • People are crap!
  • Stuff is more complicated than you think.
  • All stuff is interesting .

IMG_20150903_090148

People are crap, but not is a nasty way, we’re just not as rational as we may think.

What should we do if we want people to agree with us in a meeting? Food! Its the best predictor of acceptance of ideas. The better the food the more agreement.

Seeing is believing. Reality is not real. Dave shows example of optical illusions. We do not see absolutes, our brains fill in the gaps between reality and perception. It seems we shouldn’t trust our eyes. This can be explained by the construction of our eyes.

IMG_20150903_091602

This is not limited to only visual stimuli. Dave plays some sounds that our brains convert to words after linking them to a real sentence. Again this demonstrates the extent that our brains fill in the gaps.

Dave moves on to talk about the perception of time. How much time do you have to react to a tennis serve? A very short amount of time when you take into account the time needed to recognise the ball, the input to reach our eyes and our brain to process. It works out to approx 315ms.

Speed of serve

Fast an slow thinking allows us to only process immediately what is needed and what is easy. Fast thinking is a conditioned reflex, we make choices in our lives based on our experiences, which makes up fast thinking (system 1). System on thinking uses much less of our brain, therefore is much more resource efficient.

The problems relate back to a number of biases:

  • Confirmation Bias
  • Preference for early information
  • Group Conformity

When the waterfall model was first published it was done as proof of something that would not work. Yet it has been popular in many industries until recently.

Agile to the rescue!

So whats the point?

Science is the best invention of man. It allows us to explain the world around us. Dave thinks that Richard Feynman is one of the greatest thinkers ever.

The scientific method provides a set of guide lines that allow for repeatable experimentation. By applying the scientific method to everyday life we end up near to Agile Methodology.

I believe that this nation should commit itself to achieving the goal, before this decade is out, of landing a man on the moon and returning him safely to the earth.

The first moon landing was one of mankind’s biggest experiment. Much of the technology had not been invented yet. The problem was broken down into small chunks, was this the first Agile Project.

The Ranger programme was a series of failures until the 7th attempt.

In the marshmallow tower experiment, kindergarteners come behind only architects and CEOs. This is because they iterate on a small structure rather than plan a larger one. Another example of how experimentation works.

Dr, W. Edwards Demming applied the scientific method to a model of thinking that can be used as a framework for thinking. The ‘Toyota Improvement Kata’ is another systematic, scientific pattern of thinking. It success was in its encouragement of experimentation.

The lean mindset allows for fast and quality delivery. It eliminates waste and amplifies learning. Empowers the team and develops skills. These are all things that many businesses need and want within their organisations. These became the basis of the continuos delivery mindset.

Dave gives examples of how small errors in software have caused catastrophic effects. Many of these are due to rounding errors and incompatible calculations. If only they had unit tests. When a rounding error was discovered and fixed on the Vancouver stock exchange it almost doubled the value of its stocks.

Shorten cycle time > be experimental > build better software

  • Question everything
  • Work iteratively
  • Ask how can I test this?
  • How do I apply the scientific method?