Tennessee Leeuwenburg’s keynote on Sunday morning was about sharing and collaborating, communicating and working with non-developers, but also about getting the best out of the people you work with in general. Themes: bringing things together without needing to overlap. Charts are a universal way to communicate.
Global cost of debugging is $312 billion annually – this is 5 times the market value of Facebook. [See: press release from Cambridge University].
Be more transparent with what you put in front of people to gain credibility.
Major advantages in software (compared to other engineering) persist, if you can figure out what you want early enough.
Myth that early decisions are cheaper. They might be, but you lack visibility.
Leaders are prepared to be different. Don’t believe everything you think. Don’t get stuck in a rut.
Being a manager is like being a chef – you put the whole meal together.
Dunning-Kruger effect, “if you think you are any good at anything, read it cover to cover. Four times”. In order to assess competence, you have to have competence.
Antidote to a sense of self confidence, the Dan Pink Drive RSA Animate video.
- Mechanical skill: better pay = better performance.
- Requires at least rudimentary cognitive skill: better pay = worse performance.
- Replicated over, and over.
How do we design teams around Autonomy, Mastery, Purpose?
Need to give customers some control, but retain control.
Information introduced whilst building is annoying. More likely to accept information when “done”.
Groups of people working together will engaged in a coordinated matter. If not synchronised (not communicating) it will be worse. Disconnects lead to this situation.
- Delusion of competence.
- Technical manager.
- Domain expect manager.
Productivity of teams:
- Diversity is a solution to local minima.
- Development of trust.
- Performing as a group.
- Getting in the “idea elevator”.
- Equality of access to information.
- How to divide up thinking amongst a group.
- Individual productivity and flow state.
Transparency at all costs.
It’s about credibility and reputation.
- Forecast IO: Can see forcasts from history.
- PyPy: Code speed, see effect of checkin on speed.
- iPython notebook: Validate what is happening is real.
Convert to outcomes:
I don’t think this talk had the strongest storyline to it, but I think the themes of transparency and communication are helpful. I didn’t really see quite how relevant the discussion about Drive (Amazon) was to it – great book, and one a lot of people I knew were talking about… 4 years ago.
My main takeaway was about how to listen well enough that you can communicate with people in a way that best makes sense to them – speaking to their priorities, and values, and strengths.
One reply on “Pycon AU: Solving problems by sharing them… with Python!”
RT @catehstn: Notes from Pycon AU: Solving problems by sharing them… with Python! http://t.co/OgTPLslcKX