I had a good chat with Vinay Gupta (Vinay is organising an event about collaboration on 1st June in Bethnal Green) on Friday about what’s wrong with hackdays. Most people seem to think the hackday model is getting less effective as it’s more widely adopted. His explanation is that hack days cannot be organised top down, and have to be more spontaneous, self-organising phenomena. I’m inclined to agree that events that happen this way are more likely to succeed. This got me thinking about the way events are structure to reflect broader political views.
There are lots of experiments with non-hierarchical, fluid structures – hackdays, the Pirate Party, The Occupy Movement. The codesign events at Lambeth Council that I have been going to also fit this category – these events are basically hackdays where the participants are anyone from the community, not just developers. Software companies Github and Valve both have ‘leaderless’ internal structures.
These are examples of prefigurative politics – the idea that an organisation should reflect internally the change it wants to project on the world. For example, if you are campaigning for gender equality you must first make sure your organisation doesn’t adopt the patriarchal norms of wider society. Valve wants to disrupt and democratise the gaming industry, so it should be a democratic organisation that is open to disruption.
Applying this principle to tech is often taken to mean that your organisation should reflect the authority-less model of the Internet, exactly as Valve has interpreted it. At Lambeth Council’s codesign events that the teams are fluid, no one is formally in charge, the only rule is that you should break the rules, etc. etc. It’s collaboration inspired by the “open source model”.
The idea is to port the dynamism of the web into local government, to embrace politics of the Internet and see where it takes you.
I’ve noticed some problems with this approach at codesign events:
- Being creative is hard without constraints. There is an idea that by not specifying a tight brief the teams are liberated to explore. But actually having a totally blank canvass is incredibly oppressive, narrowing the field of options makes it much easier to be creative. This is encapsulated by the irony that telling people they must break the rules means rule breaking is impossible, since there are no rules. The New Economic Foundation’s report on codesign explicitly commends an “entirely blank sheet of paper”. For me personally this is unhelpful.
- There is still a leader. Someone always ends up as the focus of the group. Even if they refuse to be called a leader and try to be impartial they still shape the course of events. At the very least someone has to be there to insist that there is no leader, which is in itself a decision. It’s more honest to be open about this and allow participants to evaluate whether the values being imposed. If we are explicit about who the leader is, we can choose them carefully. If we are not, choice of leader is a matter of accident.
As a result, the coodesign process is often quite frustrating. Contrast this with Valve, who practice an extreme version of this approach and are highly successful. Employees have almost no guidance from above. They have wheeled desks which they roll around the building forming ad-hoc teams to achieve what they can. In this podcast, Valve’s economist-in-residence explains how he thinks it works. One prominent factor in the success of the model is that all Valve’s staff are hired because they are incredibly talented, and if they don’t perform, they are brutally sacked. They are recruited collaboratively, of necessity they are like-minded people.
This is clearly not appropriate for any organisation seeking to empower a community.
Just as the Valve model works with a set of like minded people working on the well defined problem of building a games distribution platform, hackdays work well with groups of like minded developers working on bounded problems or a selected data set.
At codesign events the problems at hand are less well defined, the people have mostly never met before and have profoundly different approaches. Without any kind of process or structure to help us there is no fixed point to grab hold of. Many non-developer oriented hacks I’ve been to are like this.
As the participants become more diverse and the problems more ambiguous, having an essentially unstructured event only adds to the sense of uncertainty and the hackday-type model starts to creak.
I’m not suggesting that voluntary events could ever function if they demand obedience in the same way as a normal workplace hierarchy, or that it would desirable if they could.
Unstructured collaboration obviously works on the web. There is no logical necessity that it will work in another context, for example with codesign, but mentioning it does feel slightly heretical. The prefigurative politics of tech are extremely powerful.