In the last weeks I discovered ProjectEuler (nice idea) and I needed to delve a little into algorithms. Wikipedia was of great help here, as it documents quite a bunch of them.
But, Wikipedia articles' style was... hmmm, not exactly what I needed. While I enjoyed the rich extra information, the historical references and math equations, what I was really looking for were the concrete implementations, preferably in C#.
So, that's how this idea came to light: how about Stack Overflow developing a Wikipedia-like branch, but strictly programming oriented?
An article on Atkins sieve for example could start by describing the algorithm then continue with implementations in various languages, C#, Python and so on. There are many answers on Stack Overflow which include references to this algorithm, with links to other related questions or external sites. Why not have it all together in a nice community article?
Of course, this would not be limited to algorithms. Paging SQL results or precompiling .NET assemblies could have their place as well...
Tagging could really help, if an article and a question have common tags, related articles could show in the sidebar, like for related questions. Or, a new feature could be added, to mark a question as a good candidate for an article; answers to that question (and related) would make a good start for creating the article.
What do you think?
EDIT:
Some answers below point to the jeopardy mechanic. Yes, it is possible but, IMO, it doesn't answer well the issues mentioned below.
How would the jeopardy thing work? Somebody decides it's time to turn information into wisdom, by consolidating answers to an interesting, general-enough question. He starts a question, and then compiles the best answers into a single one, structured like an article.
First and biggest issue, visibility. What differentiates this question from the others? It will get a visibility window for a few hours, and then it will only show in the related section, blurred into a dozen other questions. A "normal" question is not as proeminent as an article, visually promoted by Stack Overflow as such.
Second, information scattering. As this will only increase over time, the effort of building a coherent, covering answer to a given problem will be burried deeper and deeper and, as such, become fruitless.
A third issue would be text's structure. While the actual Q/A editor is good enough for formulating simple questions or answers, it misses the functionality to build something like an article: ToC, sections, references, etc.