The problem of deciding where the operational processes of any production model should happen is a classical one. From guilds to mass production, the analysis has been deeply studied for physical processes for decades but, for information society, specially about information technologies, we must admit that we have not yet reached the same level of science maturity.
Why am I stating this? The process of software development not only has exposed this situation, but it is continuously experiencing the most intense effects of the same problem. As long as the development process does not require to be located in the same location as the physical resources used, the developers might fulfil their tasks at any place.
But this is only one side of the coin: the specific characteristic of the task is that it is a highly intellectual activity. As such, it is very difficult to subject it to procedures and operational rules. And consequently, it is incredibly hard to manage this sort of development, not to mention managing a whole bunch of developers trying to work together.
All of these thoughts came to my mind while reading "Remote", the book from Jason Fried and David Heinemeier Hansson, founders of 37signals. In their book, they review the culture and disciplines which have allowed them to deal with a wildly distributed workforce all over the globe to develop their products.
While I was reading it, another book plenty of savoir faire about software development came to my mind: "Peopleware", from Tom DeMarco and Timothy Lister. A timeless classic indeed that I revisited time to time.
Both writings agree on the key of software management: trust. Without trust, there will not be empowered teams, nor outstanding performance, nor amazing agility to face the unexpected. And, of course, neither effective nor efficient distributed teams management, either.
Another common point they both discussed, with its own perspective each of them, its the environment where work takes place. In the case of "Peopleware", they analyzed how layout, furniture, open space and natural light impact on the performance and motivation of people. In "Remote", the analysis brings in the pros and cons of working at home.
But I think there might be a intense debate between the authors of both books about a very specific point: the social nature of software development activity. DeMarco and Lister remark this factor all over the text, although not expressing it directly. In the case of Fried and Heinemeier, they don't reject such important characteristic but, as they deal with it in a very distributed, very-low-frequency personal contact scenario, they use other type of mechanisms to keep the team together.
My 5 cents: to be effective and efficient, a setting as described by "Remote" require a minimum level of maturity from all members, both at professional and at personal level. Maybe you have hired young and talented people to boast your projects but, if they are highly physically distributed, much will relay on their personal traits and goodwill to keep rowing in the same direction, dealing with conflicts and stress under a continuous lack of reinforcing communication.
In a scenario which combines very different kinds of people (age, gender, education, experience, personalities... maturity), a physically closer setting, with much less dispersion, would be very helpful to any manager. And the more roomy, lit, free-configured the layout, the better.
I had once a team leader working in a very complex project, with a tight schedule (kind of perfect storm), but counting on a bunch of incredibly talented, very young people. It was the case that this leader was a serial smoker, and he needed to go out of the building frequently to indulge himself. He took advantage of this short periods of time to have casual chats with different members of his team undergoing through tough obstacles. He even used this informal meetings to deal with harsh conflicts between people. And much was obtained through these off-line chats than with dozens of formal reviews or follow-up meetings. Nothing similar might have happened in the case of highly distributed people working together.
