Sam’s talk on organizing the Atom standardization activity via a Wiki is quite nice.
Even if you have a desperate community of people yearning for some kind of exchange it is an amazingly complex coordination task finding creating the concrete standard. But, there are few patterns for solving this problem. I’m very impressed by the Atom story. It really looks like we have a new pattern, a more open pattern.
One example one of the patterns people use is to create standards: a small elite group of invited members gathers to create a first draft of the standard. This can work. It approach is not without risks. Even in the absense of bad actors it creates standards biased toward the roles represented by the members of the elite group. All exchange standards have buyers, sellers, and middlemen. When buyers and the middlemen design the standard it’s unlikely to be a best fit for the sellers. Of course all standards activities have to guard against the risk of a stick-up.
The negotiation literature suggests that one way to resolve messy multiparty negotiations is to rondevous around a single document. The Wiki fufills that role. The liturature also suggests having that document cared for by a neutral party. Sam ended up holding that bag. In the talk he uses the term lightning rod; I visualize that at that point in the talk his smoking hair stands up on end and smoke comes out of his ears. The classic example of this pattern was Switzerland, a landlocked country, acting as steward of the Treaty of the Sea. Sam wasn’t entirely a neutral party; but he was a sufficently minor player in the market that other parties didn’t see him as plotting a stick-up; well most of them didn’t.
One of my cartoons for open source is that it’s a substitute for standards making. That a portable body of code is a surprisingly good means get to a standardized way of solving some problem. The code acts as the common-document around which the parties can coordinate their efforts.
But I then tend to say that open source doesn’t know how to do standardization. That we haven’t yet managed to write any new standards using the work patterns of open source. That new standards tend to be written the using the old design patterns. That’s why the Atom example is so thought provoking; it looks like a first case of an open source style of working on a nearly pure standardization problem.
Sam was very close to totally open. Everybody with a stake was welcome to come on down. While that was helped by the highly fragmented and rapidly growing market this standard was targeted into I’m still impressed that it seems like it worked. That’s so cool.