Sputnik's background story is perhaps a bit stranger than that of many other open source projects. To make it simpler, I (Yuri, the author of most of Sputnik) will write it in the first person.
I am from Vladivostok, Russia, was educated in US, worked in California for a few years in the software industry, then started a Ph.D. program at UC Berkeley School of Information. I was looking for a research project in the field at the intersection of information studies, sociology of knowledge and sociology of technology, with some sort of international twist, so in 2005 I went to Rio de Janeiro, Brazil to do interviews with Brazilian software developments to understand their take on things. In late February of 2007 I went back to Rio to do another round of interviews and a bit of participant observation, this time focusing on small number of software projects based in Brazil. I started by looking closer at Lua and Kepler, which is a project (also based in Rio) dedicated to developing a web platform based on Lua. In case of Kepler, I decided to "put my hands on the dough" as Brazilians say. I started by helping with documentation. André Carregal, the leader of Kepler, soon decided, however, that we should use a wiki to manage the documentation. There was no wiki written in Lua/Kepler written at the time, so we looked at a few standard options, finding each lacking for one reason or another. I eventually lost patience and decided to try writing a simple wiki in Kepler.
So, on a Saturday afternoon in early April I sat down and some hours later the first version of the wiki was ready. Since the Brazilians about me emphasized my being Russian so much, I called it "Sputnik." That was my first experience programming in Lua, so it came out a bit ugly, but it did run, and after a few bug fixes, we started using it as a wiki for the Kepler project, and it was surprisingly well received. One of the nice things about working in a relatively small community is that you get noticed. I doubt anyone would pay attention to yet another wiki written in Python, but a wiki written in Lua brought it some attention early on, more than I expected. As a result, I almost immediately started a major refactoring effort, spending more time than I care to admit since April. I did manage to find time in between my hacking session to collect data for my dissertation, but let's say that my "participant observation" ended up being a lot more participant than is normal in ethnography. (I am currently working on a methodological paper about participant observation and will post a link here later.)
The later iterations of Sputnik benefitted tremendously from discussions with André (oh, the joys of face-to-face conversation) as well as contributions from other people, some "local" at Rio, others remote. Some submitted patches, others sent ideas, others reported bugs. (See Credits for a partial list of names.)
In August 2007 I returned to the United States, to be able to focus on my writing my dissertation, which eventually turned into a book, Coding Places. As I discovered, however, Sputnik by then had already started taking on a life of its own.