I was explaining to a friend a few days ago the thing I mention in my RC & MetaFilter post, about how I'm trying to avoid saying "community" when I might mean "constituency" or "group" or similar. And perhaps we should say "society" sometimes -- a group that shares some norms and heritage and places to talk with each other, but doesn't necessarily take responsibility for anything. And how the phrase "the open source community" is laughable.

But wait, they said, noting that in open source everyone has to operate by some shared rules, right?

Well, kinda, yes, I said, in that everyone's working with openly available code that's under an OSI-approved license. But they're in such different situations, and paid vs. unpaid is just part of it! Think about people who play the guitar. A rock star, a session musician, a music teacher, a member of a garage band, someone putting videos on YouTube as a kind of audition for stardom, a beginning student .... they all might be playing the same sheet music or tab, but they're really doing different things.

[The field guide to open source project archetypes that Open Tech Strategies and Mozilla are developing (PDF of the first edition) is an excellent framework for thinking about these different situations and how they structure open source projects' capabilities, who's in charge, and what you can expect.]

Then, yesterday, I was reading, then skimming, a deeeeeeeply domain-specific, detail-heavy blog post about how to implement something of particular interest to the author. And at the end, of course, they say that they're starting an open-source implementation, a prototype. And I felt as though I could see into the future -- this person creating a bit of an application, other people loving it, the project growing in popularity and importance to others, the creator wanting to step away and explore other stuff. And there the timeline fractures, depending on whether anyone took steps to get it under someone else's care, get a company to steward it or grow a vibrant collective around it.

There are a bunch of developers who want to do hard things the same way that mountain climbers want to climb hard mountains. And as a side effect of this sometimes they emit some open source artifacts, as pitons sunk into the mountainside, and if you want to follow the way they went, very closely, you can reuse those pitons. Which is great and useful.

But that is only a first step towards infrastructure -- towards robust, comfortable, safe, scalable systems. And I am the millionth person to complain on her blog about the asymmetry and fragility and just inherent ridiculousness of how much really important, widely-depended-on infrastructure in our industry is basically "reuse the leftover pitons from past explorers". But it is still complainworthy. I hereby complain. Complain! And there's a reason so many of us are doing things about it (as in my case recently and for the last five years).