The First /dev/flat

This article was published by James Aylett on Sunday, February 07, 2010 at 14:41.

This weekend a group of /dev/fort alumni gathered at a secret location in East London1 for what we’re calling /dev/flat: a bit like a fort, but smaller in all dimensions. So for two days we’ve been working on the projects started at /dev/forts 2 & 3, moving them forward towards the ultimate aim of a full launch.

One of the reasons this is valuable is that, with the benefit of some perspective, there are aspects of user flow, technical architecture or choice of restrictions that don’t really make sense once we’ve got back to the wider world. This means that getting a group of people together to work on something yields far better insights than individuals hacking in their basements. It’s also more social, which is part of the point of /dev/fort in the first place.

This isn’t the first time we’ve done this: a little over a year ago a few people from the first fort (Simon Willison, Nat Downe, Niqui Merret and James Aylett) got together a couple of times to hammer through just these kind of issues. Simon & Nat then took the baton to keep moving forward, closing issues, finishing or adding required features and so forth, to take Wildlife Near You from limited alpha out to its full public beta (and beyond, with new features now appearing regularly).

This time round we had Hannah Donovan, Steve Marshall, Mark Norman Francis and Matthew Hasler working on the /dev/fort 2 project, and George Brocklehurst, Richard Boulton and James Aylett working on the /dev/fort 3 project. One thing we discovered on both was a slight angle between the direction we’d been heading in during work at the fort, and the direction that makes sense in the cold light of day. So a lot of time was spent looking at the problem afresh (Hannah spent a couple of hours using herself as a guinea pig in an impromptu user testing lab she built in my kitchen), and then starting to rebuild the pieces that don’t make as much sense as we want.

What’s become obvious as we’ve run a number of /dev/forts is that a fort is a very good place to start projects: the isolation and lack of distraction provides a high velocity, and you feel free to get something done that makes sense at the time, without worrying so much about real users. Thereafter, you need some work to turn that first iteration into a releasable, usable website; you refine the concept, tidy up the user flows, fix those glaring errors, and put it live somewhere for more people to play with. None of this really comes as a surprise; technical projects almost always have to go through an “ideation” and experimentation phase, which can feeds into more detailed work on design and product definition, often running alongside technical work to either shore up work done in earlier experiments, or to lay the groundwork for a more robust system used from that point on.

Forts are for prototype; flats are for product.